data-raw/summary_stats.R
In yujiex/pubPriCmp: Compare Public Private Sector Energy Behavior Under Extreme Weather
library("dplyr")
library("readr")
library("pipeR")
library("feather")
library("ggplot2")
library("readxl")
library("xtable")
library("stringi")
library("zoo")
library("MASS")
library("kableExtra")
load("~/Dropbox/thesis/code/pubPriCmp/data/allData.rda")
load("~/Dropbox/thesis/code/pubPriCmp/data/allDataGreen.rda")
## EDA on buildings in East North Central region GSA portfolio
allDataGreen %>%
dplyr::select(`Name`, `Organization`, `eui_total`, `year`) %>%
dplyr::group_by(`Organization`, `Name`, `year`) %>%
dplyr::summarise(`eui_total`=sum(`eui_total`)) %>%
dplyr::ungroup() %>%
dplyr::group_by(`Organization`, `Name`) %>%
dplyr::summarise(`mean_annual_eui_total`=mean(`eui_total`)) %>%
dplyr::ungroup() %>%
ggplot2::ggplot() +
## ggplot2::geom_boxplot(ggplot2::aes(y=`mean_annual_eui_total`, x=`Organization`)) +
## ggplot2::ylim(c(0, 500)) +
ggplot2::geom_density(ggplot2::aes(x=`mean_annual_eui_total`, fill=`Organization`)) +
ggplot2::facet_wrap(.~`Organization`, ncol=1) +
ggplot2::theme()
allDataGreen %>%
dplyr::select(`Name`, `Organization`, `eui_total`, `year`, `month`) %>%
dplyr::group_by(`Organization`, `Name`, `year`) %>%
dplyr::summarise(`eui_total`=sum(`eui_total`)) %>%
dplyr::ungroup() %>%
dplyr::group_by(`Organization`, `Name`) %>%
dplyr::summarise(`mean_annual_eui_total`=mean(`eui_total`)) %>%
dplyr::ungroup() %>%
dplyr::filter(`mean_annual_eui_total` > 1000)
allDataGreen %>%
dplyr::select(`Name`, `Organization`, `eui_elec`, `eui_gas`, `year`, `month`, `GSF`) %>%
dplyr::filter(Name == "DC0001ZZ")
allDataGreen %>%
dplyr::distinct(`Name`, `GSF`) %>%
summary()
eastNorthCentral = allData %>%
dplyr::filter(`US Climate Region`=="East North Central") %>%
{.}
## plot energy trend
eastNorthCentral %>%
dplyr::mutate(`Date`=as.Date(sprintf("%s-%s-01", year, month), "%Y-%m-%d")) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
ggplot2::ggplot(ggplot2::aes(x=`Date`, y=`eui_elec_ln`, color=`Organization`, group=`Date`)) +
ggplot2::geom_violin() +
ggplot2::facet_wrap(.~`Organization`, ncol=1) +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.position = "bottom")
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/eastNorthCentral_distribution_by_time.png", width=6, height=4, units="in")
df_mean_avgminmax = feather::read_feather("building_mean_avgminmax.feather")
## plot public vs private portfolio average consumption by month
allData %>%
dplyr::filter(`Organization`=="GSA") %>%
dplyr::group_by(`Name`) %>%
dplyr::filter(sum(`>90`) > 0) %>%
dplyr::ungroup() %>%
dplyr::group_by(`Organization`, `Name`, `month`) %>%
dplyr::summarise(`eui_elec_ln`=mean(`eui_elec_ln`)) %>%
dplyr::ungroup() %>%
ggplot2::ggplot(ggplot2::aes(x=`month`, y=`eui_elec_ln`, color=`Organization`, group=`month`)) +
## ggplot2::geom_line() +
## ggplot2::geom_line() +
ggplot2::geom_boxplot() +
ggplot2::facet_wrap(.~`Organization`) +
ggplot2::theme_bw()
## plot individual building kernel smooth energy given mean monthly temperature
set.seed(0)
ratio = 0.2
number = 40
region = "East North Central"
## region = "Central"
## region = "Northeast"
## region = "Southeast"
samples =
allData %>%
dplyr::filter(`US Climate Region`==region) %>%
dplyr::select(`Name`, `Organization`) %>%
distinct(`Organization`, `Name`) %>%
dplyr::group_by(`Organization`) %>%
dplyr::slice(sample(1:n(), 30, replace=FALSE)) %>%
## dplyr::slice(sample(1:n(), as.integer(n()*ratio))) %>%
dplyr::ungroup() %>%
.$Name
## plot for all building
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::filter(`US Climate Region`==region) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
dplyr::group_by(`Name`) %>%
dplyr::mutate(`total_elec_eui`=sum(`eui_elec_ln`)) %>%
dplyr::ungroup() %>%
ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`, color=`total_elec_eui`)) +
## ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`)) +
ggplot2::geom_point(size=0.25) +
ggplot2::geom_smooth(span=1.0, se=FALSE, size=0.5) +
ggplot2::facet_wrap(~`Organization`) +
ggplot2::xlab("monthly average of daily mean temperature") +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::ggtitle(sprintf("kernel smoothing of building electricity (kBtu)/sqft on monthly mean temperature (%s)",
region)) +
ggplot2::theme_bw()
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_smooth_energy_avgtemp_all.png", gsub(" ", "", region)), width=6, height=4, units="in")
## plot a random sample of buildings
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::filter(`Name` %in% samples) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
dplyr::group_by(`Name`) %>%
dplyr::mutate(`total electricity per sqft`=sum(`eui_elec_ln`)) %>%
dplyr::ungroup() %>%
## ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`)) +
ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`, color=`total_elec_eui`)) +
ggplot2::geom_point(size=0.25) +
ggplot2::geom_smooth(span=1.0, se=FALSE, size=0.5) +
ggplot2::facet_wrap(~`Organization`) +
ggplot2::xlab("monthly average of daily mean temperature") +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::ggtitle(sprintf("kernel smoothing of building electricity (kBtu)/sqft on monthly mean temperature (%s)",
region)) +
ggplot2::theme_bw()
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_smooth_energy_avgtemp_sample.png", gsub(" ", "", region)), width=6, height=4, units="in")
## find buildings in the sample with a negative slope
downslope_buildings =
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::group_by(`Organization`, `Name`) %>%
dplyr::summarise(`slope`=lm(`eui_elec_ln`~`mean_avgminmax`)$coefficients[[2]]) %>%
dplyr::ungroup() %>%
dplyr::filter(`slope` < -0.005) %>%
.$Name
length(downslope_buildings)
## plot smoothed results for down slope buildings only
region = "East North Central"
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::filter(`Name` %in% downslope_buildings) %>%
dplyr::filter(`US Climate Region`==region) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`)) +
ggplot2::geom_point(size=0.25) +
ggplot2::geom_smooth(span=1.0, se=FALSE, size=0.5) +
ggplot2::facet_wrap(~`Organization`) +
ggplot2::xlab("monthly average of daily mean temperature") +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::ggtitle(sprintf("kernel smoothing of building electricity (kBtu)/sqft on monthly mean temperature (%s)\nwith negative slope", region)) +
ggplot2::theme_bw()
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_smooth_energy_avgtemp_downslope.png", gsub(" ", "", region)), width=6, height=4, units="in")
## find buildings in the sample with a negative slope
up_buildings =
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::group_by(`Organization`, `Name`) %>%
dplyr::summarise(`slope`=lm(`eui_elec_ln`~`mean_avgminmax`)$coefficients[[2]]) %>%
dplyr::ungroup() %>%
dplyr::filter(`slope` > 0.005) %>%
.$Name
length(up_buildings)
region = "East North Central"
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::filter(`Name` %in% up_buildings) %>%
dplyr::filter(`US Climate Region`==region) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`)) +
ggplot2::geom_point(size=0.25) +
ggplot2::geom_smooth(span=1.0, se=FALSE, size=0.5) +
ggplot2::facet_wrap(~`Organization`) +
ggplot2::xlab("monthly average of daily mean temperature") +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::ggtitle(sprintf("kernel smoothing of building electricity (kBtu)/sqft on monthly mean temperature (%s)\nwith negative slope", region)) +
ggplot2::theme_bw()
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_smooth_energy_avgtemp_up.png", gsub(" ", "", region)), width=6, height=4, units="in")
overall =
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::select(`Name`, `Organization`, `eui_elec_ln`, `<10`, starts_with("["), `>90`) %>%
{.}
buildings = unique(overall$Name)
coef_labels = c("<10", sprintf("[%s-%s)", seq(10, 80, by=10), seq(20, 90, by=10)), ">90")
coef_labels <- coef_labels[which(coef_labels != "[60-70)")]
## transpose a data frame using the first column as the new header
transpose_df_header_1st_col <- function(df, colname) {
transposed = setNames(data.frame(t(df[,-1])), df[,1][, colname]) %>%
{.}
return(transposed)
}
acc = NULL
for (b in buildings) {
print(b)
df <- overall %>%
dplyr::filter(Name == b) %>%
## dplyr::mutate(`eui_elec_ln_scaled` = `eui_elec_ln` * 1e3) %>%
{.}
out <- lm(`eui_elec_ln`~`<10` + `[10-20)` + `[20-30)` + `[30-40)` +
## out <- lm(`eui_elec_ln_scaled`~`<10` + `[10-20)` + `[20-30)` + `[30-40)` +
`[40-50)` + `[50-60)` + `[70-80)` + `[80-90)` + `>90`, data=df)
coefs = as.numeric(out$coefficients)
coefs = coefs[2:length(coefs)]
names(coefs) = coef_labels
dfcoef = data.frame(coefs) %>%
tibble::rownames_to_column() %>%
dplyr::bind_rows(data.frame(rowname="[60-70)", coefs=NA)) %>%
dplyr::mutate(`Name`=b) %>%
{.}
acc <- rbind(acc, dfcoef)
}
acc %>%
readr::write_csv("bin_coef_by_building_scaleUp1e3.csv")
org_lookup = allData %>%
distinct(`Organization`, `US Climate Region`, `Name`)
total_eui_elec_lookup = allData %>%
dplyr::group_by(`Name`) %>%
dplyr::summarise(`total electricity per sqft`=sum(`eui_elec`)) %>%
dplyr::ungroup()
acc %>%
dplyr::filter(`rowname`=="<10", coefs < -2) %>%
## dplyr::filter(`rowname`=="<10", coefs < -2000) %>%
head()
## plot coefficient estimate for individual buildings for each region
acc %>%
dplyr::mutate(`ordering`=match(`rowname`, breaklabels)) %>%
dplyr::left_join(org_lookup) %>%
dplyr::left_join(total_eui_elec_lookup) %>%
dplyr::mutate(`US Climate Region`=factor(`US Climate Region`, levels=c("Central", "East North Central", "Northeast", "Southeast", "South"))) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
## dplyr::filter(`US Climate Region`=="East North Central") %>%
## dplyr::filter(`Name` != "MI1879ZZ") %>%
ggplot2::ggplot(ggplot2::aes(x=ordering, y=coefs, group=Name, color=`total electricity per sqft`)) +
ggplot2::geom_point(size=0.5) +
ggplot2::geom_line(size=0.5) +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::xlab("Temperature bin") +
ggplot2::scale_x_continuous(breaks=1:10, labels=breaklabels) +
ggplot2::facet_wrap(`Organization`~`US Climate Region`, nrow=2) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = ggplot2::element_text(angle=90, vjust=1),
legend.position = "bottom")
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/facet_building_bin_coef.png", width=8, height=6, units="in")
## restrict the plot range to -5000 to 5000
acc %>%
dplyr::mutate(`ordering`=match(`rowname`, breaklabels)) %>%
dplyr::left_join(org_lookup) %>%
dplyr::left_join(total_eui_elec_lookup) %>%
dplyr::mutate(`US Climate Region`=factor(`US Climate Region`, levels=c("Central", "East North Central", "Northeast", "Southeast", "South"))) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
dplyr::filter(coefs < 5, coefs > -5) %>%
## dplyr::filter(`total electricity per sqft`>100) %>%
ggplot2::ggplot(ggplot2::aes(x=ordering, y=coefs, group=Name, color=`total electricity per sqft`)) +
ggplot2::geom_point(size=0.5) +
ggplot2::geom_line(size=0.5) +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::xlab("Temperature bin") +
ggplot2::scale_x_continuous(breaks=1:10, labels=breaklabels) +
ggplot2::facet_wrap(`Organization`~`US Climate Region`, nrow=2) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = ggplot2::element_text(angle=90, vjust=1),
legend.position = "bottom")
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/facet_building_bin_coef_restrictYrange.png", width=8, height=6, units="in")
## count of days in 90 bin
num_days_in_bin_90 <- allData %>%
dplyr::select(`US Climate Region`, Name, `Organization`, `eui_elec_ln`, `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`temperature bin`, `count`, `<10`:`>90`) %>%
dplyr::filter(`temperature bin` %in% c(">90")) %>%
dplyr::filter(count > 0) %>%
dplyr::group_by(`US Climate Region`, `Organization`, `temperature bin`) %>%
dplyr::summarise(count = sum(count)) %>%
dplyr::ungroup() %>%
{.}
result_table_tex <- xtable(num_days_in_bin_90, caption="Count of days with daily temperature greater than 90")
print(result_table_tex, tabular.environment = "longtable", include.rownames=FALSE,
file="~/Dropbox/thesis/writeups/policy_cmp/tables/num_days_in_bin_90.tex")
head(num_days_in_bin_90)
## plot energy against weather for each region
## region = "East North Central"
## region = "Central"
## region = "Northeast"
## region = "Southeast"
region=""
to_plot = allData
if (nchar(region) > 0) {
to_plot = allData %>%
dplyr::filter(`US Climate Region`==region) %>%
{.}
}
to_plot %>%
dplyr::select(Name, `Organization`, `eui_elec_ln`, `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`temperature bin`, `count`, `<10`:`>90`) %>%
dplyr::filter(`temperature bin` %in% c("[80-90)", ">90")) %>%
ggplot2::ggplot(ggplot2::aes(x=`count`, y=`eui_elec_ln`, color=`Organization`)) +
ggplot2::geom_jitter(size=0.5, width=0.25) +
ggplot2::facet_wrap(`Organization`~`temperature bin`) +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::xlab("number of times per month daily mean temperature fall into this bin") +
ggplot2::ggtitle("Distribution of electricity use intensity by days in high temperature bins") +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_jitter_by_bin_cnt.png", gsub(" ", "", region)), width=6, height=4, units="in")
## plot energy distribution for bin 90
allData %>%
dplyr::select(`Name`, `US Climate Region`, `Organization`, `eui_elec_ln`, `>90`) %>%
dplyr::mutate(`US Climate Region`=factor(`US Climate Region`, levels=c("Central", "East North Central", "Northeast", "Southeast", "South"))) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
ggplot2::ggplot(ggplot2::aes(x=`>90`, y=`eui_elec_ln`, color=`Organization`)) +
ggplot2::geom_jitter(size=0.5, width=0.25) +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::xlab("number of times per month daily mean temperature > 90F") +
ggplot2::facet_wrap(`Organization`~`US Climate Region`, nrow=2) +
ggplot2::theme_bw() +
ggplot2::theme(legend.position = "bottom")
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/facet_energy_distribution_by_90bin.png",
width=6, height=4, units="in")
eastNorthCentral %>%
dplyr::group_by(`Organization`, `Name`) %>%
dplyr::slice(1) %>%
ungroup() %>%
dplyr::group_by(`Organization`) %>%
dplyr::summarise(n())
## plot histogram using full data
acc = feather::read_feather("~/Dropbox/thesis/code/pubPriCmp/data-raw/all_weather.feather")
## produce histogram of daily average temperature
lowerbound = min(acc$AVGMINMAX)
upperbound = max(acc$AVGMINMAX)
breaks = c(lowerbound, seq(10, 90, by=10), upperbound)
break_labels = c("<10", sprintf("[%s-%s)", seq(10, 80, by=10), seq(20, 90, by=10)), ">90")
## produce bin count plot
acc %>%
ggplot2::ggplot(ggplot2::aes(x=`AVGMINMAX`)) +
ggplot2::geom_histogram(breaks = breaks) +
ggplot2::stat_bin(breaks = breaks, geom="text", ggplot2::aes(label = ..count..), size=2, vjust=-1) +
ggtitle("Distribution of daily mean (average of min and max) temperature") +
xlab("daily mean temperature (average of min and max)") +
scale_x_continuous(breaks=seq(-30, 100, 10)) +
theme_bw()
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/avgminmax_bin.png", width=6, height=4, units="in")
## produce bar style count of daily average temp, as in Barreca et al. 2016
acc %>%
dplyr::mutate(`value_label`=dplyr::case_when(`AVGMINMAX` < 10 ~ "<10",
`AVGMINMAX` < 20 ~ break_labels[2],
`AVGMINMAX` < 30 ~ break_labels[3],
`AVGMINMAX` < 40 ~ break_labels[4],
`AVGMINMAX` < 50 ~ break_labels[5],
`AVGMINMAX` < 60 ~ break_labels[6],
`AVGMINMAX` < 70 ~ break_labels[7],
`AVGMINMAX` < 80 ~ break_labels[8],
`AVGMINMAX` < 90 ~ break_labels[9],
TRUE ~ break_labels[10]
)) %>%
dplyr::mutate(`value_label`=factor(`value_label`, levels=break_labels)) %>%
ggplot2::ggplot(ggplot2::aes(x=`value_label`)) +
ggplot2::geom_bar() +
ggtitle("Distribution of daily mean (average of min and max) temperature,\nof the three years") +
xlab("daily mean temperature (average of min and max)") +
ggplot2::theme_bw()
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/avgminmax_bin_barstyle.png", width=6, height=4, units="in")
## summary stats of weather data
devtools::load_all("~/Dropbox/thesis/code/summaryUtil")
summaryUtil::summary_table(acc) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_weather.csv")
acc_distance = feather::read_feather("~/Dropbox/thesis/code/pubPriCmp/data-raw/station_distance.feather")
## print number of distinct weather stations used in each year
acc_distance %>%
dplyr::group_by(`start_time`) %>%
dplyr::summarise(cnt = length(unique(`id`))) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/distinct_station_cnt_byyear.csv")
## print number of distinct weather stations used
acc_distance %>%
distinct(`id`) %>%
nrow()
distance_summary = acc_distance %>%
dplyr::group_by(`Name`, `varname`, `start_time`) %>%
dplyr::summarise(`minimum distance` = min(`distance`), `maximum distance`=max(`distance`), `average distance`=mean(`distance`), `number of stations`= n()) %>%
{.}
distance_summary %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_station_distance.csv")
distance_summary %>%
dplyr::filter(cnt < 3) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/fewerThan3Stations.csv")
## distribution of max distance of weather stations available
distance_summary %>%
dplyr::group_by(`Name`, `varname`) %>%
dplyr::summarise(`max_distance_overall`=max(`maximum distance`)) %>%
dplyr::ungroup() %>%
ggplot(aes(x=max_distance_overall)) +
geom_histogram() +
ggtitle("Distribution of maximum distance between weather station and building") +
xlab("maximum distance") +
facet_wrap(~ varname) +
theme(plot.title = element_text(size=12))
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/station_max_distance.png", width=6, height=4, units="in")
load("~/Dropbox/thesis/code/pubPriCmp/data/allData.rda")
load("~/Dropbox/thesis/code/pubPriCmp/data/allDataGreen.rda")
## generate appendix summary table for all variables
devtools::load_all("~/Dropbox/thesis/code/summaryUtil")
summaryUtil::summary_table(allData) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_allData.csv")
## generate appendix summary table for all variables with Greeness rating included
devtools::load_all("~/Dropbox/thesis/code/summaryUtil")
encodedGreen <- allDataGreen %>%
dplyr::mutate(`eGreeness`=ifelse(`Greeness` == "Least Green", 1, ifelse(`Greeness` == "Moderate Green", 2, 3))) %>%
{.}
unique(encodedGreen$eGreeness)
summaryUtil::summary_table(encodedGreen) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_allDataGreen.csv")
names(allDataGreen)
## generate in-paper summary statistics
yearlyDataStatic <-
allDataGreen %>%
dplyr::select(-`latitude`, -`longitude`, -`month`, -`City`, -`hasRetrofitBeforeThisMonth`, -`Date`, -`eui_elec_ln`,
-`eui_gas_ln`) %>%
dplyr::group_by(`Name`, `year`) %>%
dplyr::summarise_at(vars(`State`, `GSF`, `Organization`, `region_wiki`, `private`, `Greeness`), funs(first)) %>%
## dplyr::summarise_at(vars(`State`, `GSF`, `Organization`, `US Climate Region`, `private`, `Greeness`), funs(first)) %>%
{.}
names(yearlyDataStatic)
yearlyDataOther <-
allDataGreen %>%
dplyr::select(-`latitude`, -`longitude`, -`month`, -`City`, -`hasRetrofitBeforeThisMonth`, -`Date`, -`eui_elec_ln`,
-`eui_gas_ln`, -`US Climate Region`, -`USRegion`) %>%
## -`eui_gas_ln`) %>%
dplyr::group_by(`Name`, `year`) %>%
## dplyr::summarise_at(vars(-one_of("State", "GSF", "Organization", "US Climate Region", "private", "Greeness")),
dplyr::summarise_at(vars(-one_of("State", "GSF", "Organization", "region_wiki", "private", "Greeness")),
funs(sum)) %>%
{.}
names(yearlyDataOther)
yearlyData <- yearlyDataStatic %>%
dplyr::left_join(yearlyDataOther, by=c("Name", "year")) %>%
{.}
yearlyAvgStatic <- yearlyDataStatic %>%
dplyr::select(-`year`) %>%
dplyr::group_by(`Name`) %>%
dplyr::summarise_all(funs(first)) %>%
{.}
yearlyAvgOther <- yearlyDataOther %>%
dplyr::select(-`year`) %>%
dplyr::group_by(`Name`) %>%
dplyr::summarise_all(funs(mean)) %>%
{.}
yearlyAvgData <- yearlyAvgStatic %>%
dplyr::left_join(yearlyAvgOther, by="Name") %>%
{.}
head(yearlyAvgData)
## ## check whether additional rows are added as a result of the join
## nrow(yearlyAvgStatic)
## nrow(yearlyAvgData)
## bar plot of temperature bins, fill by ownership
break_labels = c("<10", sprintf("[%s-%s)", seq(10, 80, by=10), seq(20, 90, by=10)), ">90")
yearlyAvgData %>%
dplyr::select(-one_of("Name", "State", "Organization", "GSF", "HDD2", "CDD2", "US Climate Region", "USRegion",
"region_wiki", "eui_gas",
## dplyr::select(-one_of("Name", "State", "Organization", "GSF", "HDD2", "CDD2", "US Climate Region", "eui_gas",
"eui_elec", "HDD", "CDD", "Electric_(kBtu)", "Gas_(kBtu)")) %>%
tidyr::gather(`Mean Temperature Bin`, `Average Number of Days`, `<10`:`>90`) %>%
dplyr::mutate(`Mean Temperature Bin`=factor(`Mean Temperature Bin`, levels=break_labels)) %>%
dplyr::group_by(`private`, `Mean Temperature Bin`) %>%
dplyr::summarise(`Average Number of Days` = mean(`Average Number of Days`)) %>%
dplyr::mutate(`Ownership` = ifelse(`private`, "private", "public")) %>%
ggplot2::ggplot(ggplot2::aes(x=`Mean Temperature Bin`, y=`Average Number of Days`, fill=`Ownership`)) +
ggplot2::geom_bar(position = "dodge", stat="identity") +
ggplot2::scale_fill_grey() +
ggplot2::ggtitle("Distribution of average annual daily mean temperature") +
ggplot2::ylab("Number of days per year") +
ggplot2::theme_bw()
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/dailyAvgDistribution.png", width=6, height=4, units="in")
## bar plot of temperature bins, fill by Greeness, facet by public and private
break_labels = c("<10", sprintf("[%s-%s)", seq(10, 80, by=10), seq(20, 90, by=10)), ">90")
yearlyAvgData %>%
dplyr::select(-one_of("Name", "State", "Organization", "GSF", "HDD2", "CDD2", "US Climate Region", "USRegion",
"region_wiki", "eui_gas",
## dplyr::select(-one_of("Name", "State", "Organization", "GSF", "HDD2", "CDD2", "US Climate Region", "eui_gas",
"eui_elec", "HDD", "CDD", "Electric_(kBtu)", "Gas_(kBtu)")) %>%
tidyr::gather(`Mean Temperature Bin`, `Average Number of Days`, `<10`:`>90`) %>%
dplyr::mutate(`Mean Temperature Bin`=factor(`Mean Temperature Bin`, levels=break_labels),
`Greeness`=factor(`Greeness`, levels=c("Most Green", "Moderate Green", "Least Green"))) %>%
dplyr::group_by(`private`, `Greeness`, `Mean Temperature Bin`) %>%
dplyr::summarise(`Average Number of Days` = mean(`Average Number of Days`)) %>%
dplyr::mutate(`Ownership` = ifelse(`private`, "private", "public")) %>%
ggplot2::ggplot(ggplot2::aes(x=`Mean Temperature Bin`, y=`Average Number of Days`, fill=`Greeness`)) +
ggplot2::geom_bar(position = "dodge", stat="identity") +
ggplot2::scale_fill_grey() +
ggplot2::ggtitle("Distribution of daily average temperature") +
ggplot2::ylab("Number of days per year") +
ggplot2::facet_wrap(~`Ownership`, nrow=2) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1))
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/dailyAvgDistribution_greeness.png", width=6, height=6, units="in")
## bar plot of temperature bins, fill by region, facet by public and private
break_labels = c("<10", sprintf("[%s-%s)", seq(10, 80, by=10), seq(20, 90, by=10)), ">90")
yearlyAvgData %>%
dplyr::select(-one_of("Name", "State", "Organization", "GSF", "HDD2", "CDD2",
"US Climate Region", "USRegion", "Greeness",
"eui_gas",
"eui_elec", "HDD", "CDD", "Electric_(kBtu)",
"Gas_(kBtu)")) %>%
tidyr::gather(`Mean Temperature Bin`, `Average Number of Days`, `<10`:`>90`) %>%
dplyr::mutate(`Mean Temperature Bin`=factor(`Mean Temperature Bin`, levels=break_labels)) %>%
dplyr::group_by(`private`, `region_wiki`, `Mean Temperature Bin`) %>%
dplyr::summarise(`Average Number of Days` = mean(`Average Number of Days`)) %>%
dplyr::mutate(`Ownership` = ifelse(`private`, "private", "public")) %>%
ggplot2::ggplot(ggplot2::aes(x=`Mean Temperature Bin`, y=`Average Number of Days`, fill=`region_wiki`)) +
ggplot2::geom_bar(position = "dodge", stat="identity") +
ggplot2::scale_fill_grey() +
ggplot2::ggtitle("Distribution of daily average temperature") +
ggplot2::ylab("Number of days per year") +
ggplot2::facet_wrap(~`Ownership`, nrow=2) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1))
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/dailyAvgDistribution_region_wiki.png", width=6, height=6, units="in")
yearlyAvgData_cnt =
yearlyAvgData %>%
dplyr::group_by(`private`, `US Climate Region`) %>%
dplyr::summarise(cnt = n()) %>%
dplyr::ungroup() %>%
dplyr::mutate(`private`=as.character(`private`)) %>%
{.}
## summary by portfolio and climate region
yearlyAvgData %>%
dplyr::select(-`Greeness`) %>%
dplyr::mutate(`<40` = `<10` + `[10-20)` + `[20-30)` + `[30-40)`) %>%
dplyr::select(-one_of("Name", "State", "Organization", "<10", "[10-20)", "[20-30)", "[30-40)", "[40-50)",
"[50-60)", "[60-70)", "[70-80)")) %>%
dplyr::group_by(`private`, `US Climate Region`) %>%
dplyr::summarise_all(funs(mean)) %>%
dplyr::ungroup() %>%
dplyr::select(`private`, `US Climate Region`, `Electric_(kBtu)`, `Gas_(kBtu)`, `eui_elec`, `eui_gas`, `HDD`,
`CDD`, `<40`, `[80-90)`, `>90`) %>%
dplyr::mutate(`private`=as.character(`private`)) %>%
dplyr::mutate_if(is.numeric, function(x) {sprintf("%.1f", x)}) %>%
dplyr::left_join(yearlyAvgData_cnt, by=c("private", "US Climate Region")) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_annual_avg.csv")
## summary overall
yearlyAvgData %>%
dplyr::mutate(`<40` = `<10` + `[10-20)` + `[20-30)` + `[30-40)`) %>%
dplyr::select(-one_of("Name", "State", "Organization", "<10", "[10-20)", "[20-30)", "[30-40)", "[40-50)",
"[50-60)", "[60-70)", "[70-80)", "GSF", "US Climate Region", "private", "Greeness")) %>%
dplyr::summarise_all(mean) %>%
dplyr::select(`Electric_(kBtu)`, `Gas_(kBtu)`, `eui_elec`, `eui_gas`, `HDD`,
`CDD`, `<40`, `[80-90)`, `>90`) %>%
dplyr::mutate_if(is.numeric, function(x) {sprintf("%.1f", x)}) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_annual_avg_nogroup.csv")
## summary public and private
yearlyAvgData %>%
dplyr::mutate(`<40` = `<10` + `[10-20)` + `[20-30)` + `[30-40)`) %>%
dplyr::select(-one_of("Name", "State", "Organization", "<10", "[10-20)", "[20-30)", "[30-40)", "[40-50)",
"[50-60)", "[60-70)", "[70-80)", "GSF", "US Climate Region", "Greeness")) %>%
dplyr::group_by(`private`) %>%
dplyr::summarise_all(funs(mean)) %>%
dplyr::ungroup() %>%
dplyr::select(`private`, `Electric_(kBtu)`, `Gas_(kBtu)`, `eui_elec`, `eui_gas`, `HDD`,
`CDD`, `<40`, `[80-90)`, `>90`) %>%
dplyr::mutate(`private`=as.character(`private`)) %>%
dplyr::mutate_if(is.numeric, function(x) {sprintf("%.1f", x)}) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_annual_avg_pubpri.csv")
## summary by portfolio and greeness level
yearlyAvgData_cnt =
yearlyAvgData %>%
dplyr::group_by(`private`, `Greeness`) %>%
dplyr::summarise(cnt = n()) %>%
dplyr::ungroup()
yearlyAvgData %>%
dplyr::mutate(`<40` = `<10` + `[10-20)` + `[20-30)` + `[30-40)`) %>%
dplyr::select(-`US Climate Region`) %>%
dplyr::select(-one_of("Name", "State", "Organization", "<10", "[10-20)", "[20-30)", "[30-40)", "[40-50)",
"[50-60)", "[60-70)", "[70-80)")) %>%
dplyr::group_by(`private`, `Greeness`) %>%
dplyr::summarise_all(funs(mean)) %>%
dplyr::ungroup() %>%
dplyr::select(`private`, `Greeness`, `Electric_(kBtu)`, `Gas_(kBtu)`, `eui_elec`, `eui_gas`, `HDD`,
`CDD`, `<40`, `[80-90)`, `>90`) %>%
dplyr::mutate_if(is.numeric, function(x) {sprintf("%.1f", x)}) %>%
dplyr::mutate(`private`=as.numeric(`private`)) %>%
dplyr::left_join(yearlyAvgData_cnt, by=c("private", "Greeness")) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_annual_avg_byGreeness.csv")
## produce plots of coefficients for temperature bin, with confidence interval
breaklevels = c("avgTempBelow10", "avgTemp10to20", "avgTemp20to30", "avgTemp30to40",
"avgTemp40to50", "avgTemp50to60", "avgTemp60to70", "avgTemp70to80", "avgTemp80to90",
"avgTempAbove90")
breaklabels = c("<10", "[10-20)", "[20-30)", "[30-40)",
"[40-50)", "[50-60)", "[60-70)", "[70-80)", "[80-90)",
">90")
load("../data/allData.rda")
## ------------------------------------------------------------------------------
## main analysis plots older version start
## ------------------------------------------------------------------------------
## portfolio = ""
## ## "_gas" is gas, "" is electricity
## energy_type = "_gas"
## status_code = ""
## df_zero = data.frame(X1="avgTemp60to70", variable=rep(c("coef", "ci_low", "ci_high"), 6),
## model=rep(c(rep("X2", 3), rep("X3", 3), rep("X4", 3)), 2),
## value = 0, status=c(rep("public", 9), rep("private relative to public", 9)))
## portfolio = "_GSA"
## ## energy_type = "_total"
## ## energy_type = "_gas"
## energy_type = ""
## status_code = "public"
## df_zero = data.frame(X1="avgTemp60to70", variable=rep(c("coef", "ci_low", "ci_high"), 2),
## model=rep(c(rep("X2", 3), rep("X3", 3)), 2),
## value = 0, status=c(rep("public", 6)))
portfolio = "_GSA"
## portfolio = "_PNC"
energy_type = "_total"
## energy_type = "_gas"
## energy_type = ""
## status_code = "private"
status_code = "public"
## office.class = "o_o"
## folder.suffix = paste0("_", office.class)
## office.class = ""
## folder.suffix = ""
for (portfolio in c("_GSA", "_PNC")) {
## could change this to other energy_type
for (ownership in c("owned", "owned_and_leased")) {
## energy_type = "_total"
for (energy_type in c("_gas")) {
## for (energy_type in c("", "_gas")) {
if (portfolio == "_PNC") {
status_code = "private"
} else if (portfolio == "_GSA") {
status_code = "public"
}
df_zero = data.frame(X1="avgTemp60to70", variable=rep(c("coef", "ci_low", "ci_high"), 2),
model=rep(c(rep("X2", 3), rep("X3", 3)), 2),
value = 0, status=c(rep(status_code, 6)))
for (office.class in c("", "o_o", "o_oct", "r_o", "ro_oct")) {
if (nchar(office.class) == 0) {
folder.suffix = ""
} else {
folder.suffix = paste0("_", office.class, "_", ownership)
}
## produce the bin coefficient and CI plot for the model with both portfolio
df =
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/reg_result%s/regression_results_bin_0720%s%s.txt", folder.suffix, tolower(portfolio), energy_type), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
tidyr::gather(`model`, `value`, starts_with("X")) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("0b.", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE),
!grepl("hasretrofit", `X1`, fixed = TRUE),
!grepl("Constant", `X1`, fixed = TRUE)
) %>%
{.}
if (status_code == "") {
df <- df %>%
dplyr::mutate(`status`=ifelse(grepl("1.private", X1, fixed=TRUE), "private relative to public", "public")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private relative to public"))) %>%
{.}
} else {
df <- df %>%
dplyr::mutate(`status`=status_code) %>%
{.}
}
df <- df %>%
dplyr::mutate(`X1`=gsub("1.private#c.", "", `X1`)) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
## readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/ci_bin.csv")
dplyr::bind_rows(df_zero) %>%
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="base line", "X3"="state specific trend", "X4"="state trend x private") %>%
{.}
if (nchar(portfolio) == 1) {
df <- df %>%
dplyr::filter(`model`=="state trend x private") %>%
{.}
} else {
df <- df %>%
dplyr::filter(`model`=="state specific trend") %>%
{.}
}
df <- df %>%
dplyr::filter(!(`X1` %in% c("avgTempBelow10", "avgTempAbove90"))) %>%
{.}
head(df)
min(df$value)
max(df$value)
if (energy_type == "") {
if (office.class == "") {
shifty = -8
yupperlimit = 15
}
if (ownership == "owned") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -30
yupperlimit = 40
} else if (office.class == "r_o") {
shifty = -8
yupperlimit = 15
} else if (office.class == "ro_oct") {
shifty = -8
yupperlimit = 15
}
} else if (ownership == "owned_and_leased") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -18
yupperlimit = 30
} else if (office.class == "r_o") {
shifty = -8
yupperlimit = 20
} else if (office.class == "ro_oct") {
shifty = -8
yupperlimit = 15
}
}
## shifty = -25
## yupperlimit = 25
scaling = 1 / 1e5 * 1.5
ylabel = "Electricity"
} else if (energy_type == "_gas") {
if (office.class == "") {
shifty = -10
yupperlimit = 55
}
if (ownership == "owned") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -40
yupperlimit = 75
} else if (office.class == "r_o") {
shifty = -10
yupperlimit = 70
} else if (office.class == "ro_oct") {
shifty = -10
yupperlimit = 60
}
} else if (ownership == "owned_and_leased") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -40
yupperlimit = 63
} else if (office.class == "r_o") {
shifty = -10
yupperlimit = 60
} else if (office.class == "ro_oct") {
shifty = -10
yupperlimit = 50
}
}
## shifty = -20
## yupperlimit = 40
scaling = 1 / 1e5 * 2
ylabel = "Gas"
} else if (energy_type == "_total") {
if (office.class == "") {
shifty = -10
yupperlimit = 40
}
if (ownership == "owned") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -30
yupperlimit = 60
} else if (office.class == "r_o") {
shifty = -10
yupperlimit = 45
} else if (office.class == "ro_oct") {
shifty = -10
yupperlimit = 40
}
} else if (ownership == "owned_and_leased") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -30
yupperlimit = 55
} else if (office.class == "r_o") {
shifty = -10
yupperlimit = 55
} else if (office.class == "ro_oct") {
shifty = -10
yupperlimit = 40
}
}
scaling = 1 / 1e5 * 2
ylabel = "Electricity + Gas"
}
temperatureBinData =
allData %>%
{.}
if (ownership == "owned") {
temperatureBinData <- temperatureBinData %>%
dplyr::filter(Ownership == "Owned")
}
if (office.class == "o_o") {
generic.office.pnc = "Office"
generic.office.gsa = "Office"
} else if (office.class == "o_oct") {
generic.office.pnc = "Office"
generic.office.gsa = c("Courthouse", "Office", "CT/Office")
} else if (office.class == "r_o") {
generic.office.pnc = "Retail"
generic.office.gsa = "Office"
} else if (office.class == "ro_oct") {
generic.office.pnc = c("Retail", "Office")
generic.office.gsa = c("Courthouse", "Office", "CT/Office")
}
if (office.class != "") {
temperatureBinData <- temperatureBinData %>%
dplyr::filter((Organization == "PNC" & `type_general` %in% generic.office.pnc) | (Organization == "GSA" & `type_general` %in% generic.office.gsa)) %>%
{.}
}
temperatureBinData <- temperatureBinData %>%
dplyr::select(`Organization`, `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
dplyr::group_by(`Organization`, `tempBin`) %>%
dplyr::summarise(count = sum(value)) %>%
dplyr::ungroup() %>%
dplyr::mutate(`count`=`count` * scaling) %>%
dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
{.}
if (status_code == "") {
temperatureBinData <- temperatureBinData %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private relative to public")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private relative to public"))) %>%
dplyr:::select(`count_start`, `count_end`, `ordering`, `status`) %>%
{.}
} else {
temperatureBinData <- temperatureBinData %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
dplyr:::select(`count_start`, `count_end`, `ordering`, `status`) %>%
{.}
}
head(temperatureBinData)
## dplyr::mutate_at(vars(variable), recode, "ci_low"="95% C.I. low end", "ci_high"="95% C.I. high end", "coef"="coefficient") %>%
df %>%
dplyr::left_join(temperatureBinData, by=c("status", "ordering")) %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value)) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:10, labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`, xend=`ordering`, y=`count_start`, yend=`count_end`), size=10) +
## ggplot2::geom_bar(mapping=ggplot2::aes(x=`ordering`, y=`count`), stat="identity", data=temperatureBinData) +
ggplot2::facet_wrap(`status`~`model`, ncol=1) +
ggplot2::xlab("Temperature Bin") +
ggplot2::ylab(sprintf("ln(%s in kBtu / gross sqft) * 1000", ylabel)) +
ggplot2::coord_cartesian(ylim=c(shifty, yupperlimit)) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci%s%s%s.png", portfolio, energy_type, folder.suffix), width=5, height=5, units="in")
}
}
}
}
## ------------------------------------------------------------------------------
## main analysis plots older version end
## ------------------------------------------------------------------------------
## ------------------------------------------------------------------------------
## main analysis plots newer version start
## ------------------------------------------------------------------------------
folder.suffixes = c("",
"_o_o_owned", "_o_oct_owned", "_r_o_owned", "_ro_oct_owned",
"_o_o_owned_and_leased", "_o_oct_owned_and_leased", "_r_o_owned_and_leased", "_ro_oct_owned_and_leased"
)
captions = c("no restrictions",
"Public: owned offices; Private: owned offices",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices",
"Public: owned offices; Private: owned retails",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices, retails",
"Public: offices; Private: offices",
"Public: offices, courthouses, CT/Offices; Private: offices",
"Public: offices; Private: retails",
"Public: offices, courthouses, CT/Offices; Private: offices, retails")
## energy_type = "_total"
energy_type = ""
## energy_type = "_gas"
for (modelsuffix in c("baseline", "statetrend")) {
for (i in seq_along(folder.suffixes)) {
s = folder.suffixes[i]
print(s)
tokens = unlist(strsplit(s, "_"))
if (length(tokens) == 0) {
ownlease = NA
usetype = NA
} else {
usetype = paste0(tokens[2:3], collapse = "_")
if ("leased" %in% tokens) {
ownlease = NA
} else {
ownlease = "Owned"
}
}
filename = sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/reg_result%s/regression_portfolio_%s%s.txt", s, modelsuffix,
energy_type)
fields_to_remove = c(".eDate", "eState", "Constant", "hasretrofitbeforethismonth", "0b.", "1b.eGreeness")
fields_to_keep = NULL
outreg_vars = c("coef", "se", "ci_low", "ci_high")
vars_to_keep = c("coef", "ci_low", "ci_high")
df <- read_outreg2(filename=filename, fields_to_remove= fields_to_remove, fields_to_keep= fields_to_keep,
outreg_vars=outreg_vars, vars_to_keep=vars_to_keep)
print(head(df))
model_info = tibble::tibble(`model_header`=sprintf("X%s", 2:(2 + 1)),
status=c("public", "private"))
df_zero = tibble::tibble(X1="avgTemp60to70",
variable=rep(c("coef", "ci_low", "ci_high"), 2),
model=modelsuffix,
value = 0,
status=c(rep("public", 3), rep("private", 3))) %>%
{.}
df <- df %>%
tidyr::gather(`model_header`, `value`, X2:(!!rlang::sym(sprintf("X%s", 2 + 1)))) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
dplyr::left_join(model_info, by="model_header") %>%
dplyr::mutate(`model`=modelname) %>%
dplyr::bind_rows(df_zero) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::select(-`model_header`) %>%
{.}
dfresult <- df %>%
dplyr::filter(!(`X1` %in% c("avgTempBelow10", "avgTempAbove90"))) %>%
{.}
lower = min(dfresult$value)
upper = max(dfresult$value)
temperatureBinData = get.study.set(ownlease, usetype)$df
temperatureSummary <- temperatureBinData %>%
dplyr::select(`Organization`, !!rlang::sym(regioncol), `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
dplyr::group_by(`Organization`, !!rlang::sym(regioncol), `tempBin`) %>%
dplyr::summarise(count = sum(value)) %>%
dplyr::ungroup() %>%
{.}
shifty = (-1)*(upper - lower)/2
scaling = abs(shifty) / max(temperatureSummary$count) / 2
temperatureToPlot <- temperatureSummary %>%
dplyr::mutate(`count`=`count` * scaling) %>%
dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
dplyr::select(`count_start`, `count_end`, `ordering`, `status`, !!rlang::sym(regioncol)) %>%
dplyr::rename(Rating = !!rlang::sym(regioncol)) %>%
{.}
xlabel = "Temperature Bin"
if (energy_type == "") {
ylabel = "Electricity"
} else if (energy_type == "_total") {
ylabel = "Electricity + Gas"
} else if (energy_type == "_gas") {
ylabel = "Gas"
}
## facet on public private and greeness level
dfresult %>%
dplyr::left_join(temperatureToPlot, by=c("ordering", "status")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private"))) %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value)) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`, xend=`ordering`, y=`count_start`, yend=`count_end`), size=7) +
## for slide
ggplot2::facet_wrap(.~`status`) +
## for paper
## ggplot2::facet_wrap(`Rating`~`status`, ncol=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(sprintf("ln(%s in kBtu / gross sqft) * 1000", ylabel)) +
## ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::labs(title=captions[i],
subtitle=(sprintf("%s model", modelsuffix))) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_separate_organization_%s_slide%s%s.png",
modelsuffix, energy_type, s), width=8, height=5, units="in")
}
}
## ------------------------------------------------------------------------------
## main analysis plots newer version start
## ------------------------------------------------------------------------------
breaklevels = c("hdd", "hdd2", "cdd", "cdd2")
breaklabels = c("HDD", "HDD2", "CDD", "CDD2")
xlabel = "Monthly HDD CDD"
## produce the hdd cdd model coefficient and CI plot for the model with both portfolio
readr::read_tsv("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_reg_0720.txt", skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`model`, `value`, X2:X7) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("0b.", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE),
!grepl("hasretrofit", `X1`, fixed = TRUE),
!grepl("Constant", `X1`, fixed = TRUE)
) %>%
dplyr::mutate(`status`=ifelse(grepl("1.private", X1, fixed=TRUE), "private relative to public", "public")) %>%
dplyr::mutate(`X1`=gsub("1.private#c.", "", `X1`)) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
## readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/ci_reg.csv")
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="baseline", "X3"="baseline (poly 2)", "X4" = "state specific trend",
"X5" = "state specific trend (poly 2)", "X6"="state trend x private",
"X7"="state trend x private (poly 2)") %>%
## dplyr::mutate_at(vars(variable), recode, "ci_low"="95% C.I. low end", "ci_high"="95% C.I. high end", "coef"="coefficient") %>%
dplyr::filter(`model` %in% c("state trend x private", "state trend x private (poly 2)")) %>%
tidyr::unite(`temp`, `X1`, `model`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "model"), sep="-") %>%
ggplot2::ggplot() +
ggplot2::geom_pointrange(ggplot2::aes(x=ordering, ymin=`ci_low`, ymax=`ci_high`, y=`coef`),
shape=18) +
## ggplot2::geom_errorbar(ggplot2::aes(x=ordering, ymin=`ci_low`, ymax=`ci_high`)) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::facet_wrap(`status`~`model`, nrow=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(NULL) +
ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank())
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/reg_coef_ci.png", width=5, height=5, units="in")
## portfolio = "GSA"
## status_code = "public"
portfolio = "PNC"
status_code = "private"
## produce the hdd cdd model coefficient and CI plot for the model for just one portfolio
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_reg_0720_%s.txt", portfolio), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`model`, `value`, X2:X5) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("0b.", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE),
!grepl("hasretrofit", `X1`, fixed = TRUE),
!grepl("Constant", `X1`, fixed = TRUE)
) %>%
dplyr::mutate(`status`=status_code) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
## readr::write_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/ci_reg_%s.csv", portfolio))
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="baseline", "X3"="baseline (poly 2)", "X4" = "state specific trend",
"X5" = "state specific trend (poly 2)") %>%
dplyr::filter(`model` %in% c("state specific trend", "state specific trend (poly 2)")) %>%
tidyr::unite(`temp`, `X1`, `model`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "model"), sep="-") %>%
ggplot2::ggplot() +
ggplot2::geom_pointrange(ggplot2::aes(x=ordering, ymin=`ci_low`, ymax=`ci_high`, y=`coef`), shape=18) +
## ggplot2::geom_errorbar(ggplot2::aes(x=ordering, ymin=`ci_low`, ymax=`ci_high`)) +
## ggplot2::geom_point(ggplot2::aes(x=ordering, y=coef)) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::facet_wrap(`status`~`model`, nrow=1) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(NULL) +
ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank())
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/reg_coef_ci_%s.png", portfolio), width=4, height=3, units="in")
breaklevels = c("avgTempBelow10", "avgTemp10to20", "avgTemp20to30", "avgTemp30to40",
"avgTemp40to50", "avgTemp50to60", "avgTemp60to70", "avgTemp70to80", "avgTemp80to90",
"avgTempAbove90")
breaklabels = c("<10", "[10-20)", "[20-30)", "[30-40)",
"[40-50)", "[50-60)", "[60-70)", "[70-80)", "[80-90)",
">90")
## produce coefficient plot of greeness for separate portfolio using error bars
portfolio = "GSA"
status_code = "public"
## portfolio = "PNC"
## status_code = "private"
xlabel = "Temperature Bin"
## produce the bin model coefficient and CI plot for greeness analysis for the model for just one portfolio
df <- readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_bin_greeness_0720_%s.txt", portfolio), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`model`, `value`, X2:X3) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("0b.", `X1`, fixed = TRUE),
!grepl("1b.eGreeness", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE),
!grepl("hasretrofit", `X1`, fixed = TRUE),
!grepl("Constant", `X1`, fixed = TRUE)
) %>%
dplyr::mutate(`status`=status_code) %>%
dplyr::mutate(`Rating`=ifelse(grepl("2.eGreeness", `X1`, fixed=TRUE), "Moderate Green \nrelative to Least Green",
ifelse(grepl("3.eGreeness", `X1`, fixed=TRUE), "Most Green \nrelative to Least Green", "Least Green"))) %>%
dplyr::mutate(`X1`=gsub("2.eGreeness#c.", "", `X1`)) %>%
dplyr::mutate(`X1`=gsub("3.eGreeness#c.", "", `X1`)) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="baseline", "X3" = "state specific trend") %>%
## readr::write_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/ci_bin_greeness_%s.csv", portfolio))
tidyr::unite(`temp`, `X1`, `model`, `Rating`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "model", "Rating"), sep="-") %>%
{.}
df %>%
ggplot2::ggplot() +
ggplot2::geom_errorbar(ggplot2::aes(x=ordering, ymin=`ci_low`, ymax=`ci_high`)) +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=coef)) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::facet_wrap(`Rating`~`model`, ncol=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(NULL) +
## ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank())
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness_%s.png", portfolio), width=8, height=5, units="in")
## produce coefficient plot of greeness for separate portfolio using connected dashed lines for conf int
df_zero = data.frame(X1="avgTemp60to70", variable=rep(c("coef", "ci_low", "ci_high"), 6),
model=rep(c(rep("X2", 9), rep("X3", 9)), 2),
Rating=rep(c(rep("Least Green", 3), rep("Moderate Green \nrelative to Least Green", 3),
rep("Most Green \nrelative to Least Green", 3)), 2),
value = 0, status=status_code)
## portfolio = "GSA"
## status_code = "public"
## energy_type = "_gas"
portfolio = "PNC"
status_code = "private"
energy_type = "_gas"
xlabel = "Temperature Bin"
## produce the bin model coefficient and CI plot for greeness analysis for the model for just one portfolio
df <- readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_bin_greeness_0720_%s%s.txt", portfolio, energy_type), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`model`, `value`, X2:X3) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("0b.", `X1`, fixed = TRUE),
!grepl("1b.eGreeness", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE),
!grepl("hasretrofit", `X1`, fixed = TRUE),
!grepl("Constant", `X1`, fixed = TRUE)
) %>%
dplyr::mutate(`status`=status_code) %>%
dplyr::mutate(`Rating`=ifelse(grepl("2.eGreeness", `X1`, fixed=TRUE), "Moderate Green \nrelative to Least Green",
ifelse(grepl("3.eGreeness", `X1`, fixed=TRUE), "Most Green \nrelative to Least Green", "Least Green"))) %>%
dplyr::mutate(`X1`=gsub("2.eGreeness#c.", "", `X1`)) %>%
dplyr::mutate(`X1`=gsub("3.eGreeness#c.", "", `X1`)) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
dplyr::bind_rows(df_zero) %>%
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="baseline", "X3" = "state specific trend") %>%
## readr::write_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/ci_bin_greeness_%s.csv", portfolio))
{.}
min(df$value)
max(df$value)
if (energy_type == "") {
shifty = -45
yupperlimit = 25
scaling = 1 / 1e4 * 1.5
} else if (energy_type == "_gas") {
if (status_code == "public") {
shifty = -260
} else {
shifty = -100
}
yupperlimit = 75
scaling = 1 / 1e4 * 5
}
temperatureBinData =
allDataGreen %>%
{.}
temperatureBinData <-
temperatureBinData %>%
dplyr::select(`Organization`, `Greeness`, `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
dplyr::group_by(`Organization`, `Greeness`, `tempBin`) %>%
dplyr::summarise(count = sum(value)) %>%
dplyr::ungroup() %>%
dplyr::mutate(`count`=`count` * scaling) %>%
dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
{.}
if (status_code == "") {
temperatureBinData <- temperatureBinData %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private relative to public")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private relative to public"))) %>%
dplyr:::select(`count_start`, `count_end`, `ordering`, `status`, `Greeness`) %>%
{.}
} else {
temperatureBinData <- temperatureBinData %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
dplyr:::select(`count_start`, `count_end`, `ordering`, `status`, `Greeness`) %>%
{.}
}
temperatureBinData <- temperatureBinData %>%
dplyr::mutate(`Rating`=recode(Greeness, "Moderate Green"="Moderate Green \nrelative to Least Green",
"Most Green"="Most Green \nrelative to Least Green")) %>%
dplyr::select(-`Greeness`) %>%
{.}
df %>%
dplyr::left_join(temperatureBinData, by=c("ordering", "status", "Rating")) %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value)) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`, xend=`ordering`, y=`count_start`, yend=`count_end`), size=7) +
## for slide
ggplot2::facet_wrap(`model`~`Rating`, ncol=3) +
## for paper
## ggplot2::facet_wrap(`Rating`~`model`, ncol=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab("ln(Electricity in kBtu / gross sqft) * 1000") +
## ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness_%s_slide%s.png", portfolio, energy_type), width=8, height=5, units="in")
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness_%s.png", portfolio), width=6, height=7, units="in")
## produce coefficient plot of greeness for separate portfolio by greeness level
## using connected dashed lines for conf int
read_outreg2 <- function(filename, fields_to_remove, fields_to_keep, outreg_vars,
vars_to_keep) {
df =
readr::read_tsv(filename, skip=7, col_names=FALSE) %>%
head(-5) %>%
dplyr::mutate(`variable`=rep(outreg_vars, nrow(.)/length(outreg_vars))) %>%
tidyr::fill(`X1`) %>%
dplyr::filter(`variable` %in% vars_to_keep) %>%
{.}
for (field in fields_to_remove) {
df <- df %>%
dplyr::filter(!grepl(field, `X1`, fixed = TRUE)) %>%
{.}
}
for (field in fields_to_keep) {
df <- df %>%
dplyr::filter(grepl(field, `X1`, fixed = TRUE)) %>%
{.}
}
return(df)
}
breaklevels = c("avgTempBelow10", "avgTemp10to20", "avgTemp20to30", "avgTemp30to40",
"avgTemp40to50", "avgTemp50to60", "avgTemp60to70", "avgTemp70to80", "avgTemp80to90",
"avgTempAbove90")
breaklabels = c("<10", "[10-20)", "[20-30)", "[30-40)",
"[40-50)", "[50-60)", "[60-70)", "[70-80)", "[80-90)",
">90")
## green_levels = c("Least Green", "Moderate Green", "Most Green")
green_levels = c("Least Green", "Most Green")
len_green_levels = length(green_levels)
## energy_type = "_gas"
## modelsuffix = "baseline"
## modelname = "baseline"
## energy_type = "_gas"
energy_type = "_total"
modelsuffix = "statetrend"
modelname = "state specific trend"
## regioncol = "Greeness"
## filenamesuffix = ""
## green.source = "Majority Vote"
## regioncol = "greeness.howe.state"
## filenamesuffix = "_howestate"
## green.source = "State Data from Howe 2015"
regioncol = "greeness.howe.county"
filenamesuffix = "_howecounty"
green.source = "County Data from Howe 2015"
folder.suffixes = c("",
"_o_o_owned", "_o_oct_owned", "_r_o_owned", "_ro_oct_owned",
"_o_o_owned_and_leased", "_o_oct_owned_and_leased", "_r_o_owned_and_leased", "_ro_oct_owned_and_leased"
)
captions = c("no restrictions",
"Public: owned offices; Private: owned offices",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices",
"Public: owned offices; Private: owned retails",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices, retails",
"Public: offices; Private: offices",
"Public: offices, courthouses, CT/Offices; Private: offices",
"Public: offices; Private: retails",
"Public: offices, courthouses, CT/Offices; Private: offices, retails")
for (i in seq_along(folder.suffixes)) {
s = folder.suffixes[i]
print(s)
tokens = unlist(strsplit(s, "_"))
if (length(tokens) == 0) {
ownlease = NA
usetype = NA
} else {
usetype = paste0(tokens[2:3], collapse = "_")
if ("leased" %in% tokens) {
ownlease = NA
} else {
ownlease = "Owned"
}
}
filename = sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/reg_result%s/regression_greeness_%s%s%s.txt", s, modelsuffix,
energy_type, filenamesuffix)
fields_to_remove = c(".eDate", "eState", "Constant", "hasretrofitbeforethismonth", "0b.", "1b.eGreeness")
fields_to_keep = NULL
outreg_vars = c("coef", "se", "ci_low", "ci_high")
vars_to_keep = c("coef", "ci_low", "ci_high")
df <- read_outreg2(filename=filename, fields_to_remove= fields_to_remove, fields_to_keep= fields_to_keep,
outreg_vars=outreg_vars, vars_to_keep=vars_to_keep)
head(df)
model_info = tibble::tibble(`model_header`=sprintf("X%s", 2:(2*len_green_levels + 1)),
status=c(rep("public", len_green_levels), rep("private", len_green_levels)),
Rating=rep(green_levels, 2))
df_zero = tibble::tibble(X1="avgTemp60to70",
variable=rep(c("coef", "ci_low", "ci_high"), 2 * len_green_levels),
model=modelname,
Rating=rep(sort(rep(green_levels, 3)), 2), value = 0,
status=c(rep("public", 3 * len_green_levels), rep("private", 3 * len_green_levels))) %>%
{.}
df <- df %>%
tidyr::gather(`model_header`, `value`, X2:(!!rlang::sym(sprintf("X%s", 2 * len_green_levels + 1)))) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
dplyr::left_join(model_info, by="model_header") %>%
dplyr::mutate(`model`=modelname) %>%
dplyr::bind_rows(df_zero) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::select(-`model_header`) %>%
{.}
dfresult <- df %>%
dplyr::filter(!(`X1` %in% c("avgTempBelow10", "avgTempAbove90"))) %>%
{.}
lower = min(dfresult$value)
upper = max(dfresult$value)
temperatureBinData = get.study.set(ownlease, usetype)$df
temperatureSummary <- temperatureBinData %>%
dplyr::select(`Organization`, !!rlang::sym(regioncol), `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
dplyr::group_by(`Organization`, !!rlang::sym(regioncol), `tempBin`) %>%
dplyr::summarise(count = sum(value)) %>%
dplyr::ungroup() %>%
{.}
shifty = (-1)*(upper - lower)/2
scaling = abs(shifty) / max(temperatureSummary$count) / 2
temperatureToPlot <- temperatureSummary %>%
dplyr::mutate(`count`=`count` * scaling) %>%
dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
dplyr::select(`count_start`, `count_end`, `ordering`, `status`, !!rlang::sym(regioncol)) %>%
dplyr::rename(Rating = !!rlang::sym(regioncol)) %>%
{.}
## if (energy_type == "") {
## shifty = -45
## yupperlimit = 25
## scaling = 1 / 1e4 * 1.5
## ylabel = "Electricity"
## } else if (energy_type == "_gas") {
## shifty = -45
## yupperlimit = 240
## scaling = 1 / 1e4 * 1.5
## ylabel = "Gas"
## } else if (energy_type == "_total") {
## shifty = -5
## yupperlimit = 40
## scaling = 1 / 1e5 * 1.5
## ylabel = "Electricity + Gas"
## }
## temperatureBinData =
## allDataGreen %>%
## {.}
## temperatureBinData <-
## temperatureBinData %>%
## dplyr::select(`Organization`, `Greeness`, `<10`, starts_with("["), `>90`) %>%
## tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
## dplyr::group_by(`Organization`, `Greeness`, `tempBin`) %>%
## dplyr::summarise(count = sum(value)) %>%
## dplyr::ungroup() %>%
## dplyr::mutate(`count`=`count` * scaling) %>%
## dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
## dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
## dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
## dplyr::select(`count_start`, `count_end`, `ordering`, `status`, `Greeness`) %>%
## dplyr::rename(`Rating`=`Greeness`) %>%
## {.}
## head(temperatureBinData)
## head(df)
xlabel = "Temperature Bin"
ylabel = "Electricity + Gas"
## facet on public private and greeness level
dfresult %>%
dplyr::left_join(temperatureToPlot, by=c("ordering", "status", "Rating")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private"))) %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value)) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`, xend=`ordering`, y=`count_start`, yend=`count_end`), size=7) +
## for slide
ggplot2::facet_wrap(`status`~`Rating`, ncol=len_green_levels) +
## for paper
## ggplot2::facet_wrap(`Rating`~`status`, ncol=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(sprintf("ln(%s in kBtu / gross sqft) * 1000", ylabel)) +
## ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::labs(title=captions[i],
subtitle=(sprintf("%s model (greeness source: %s)", modelsuffix, green.source))) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness%s_separate_organization_%s_slide%s%s.png",
filenamesuffix, modelsuffix, energy_type, s), width=8, height=5, units="in")
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness_separate_organization_greeness_%s.png", modelsuffix), width=6, height=7, units="in")
## facet on public private and fill on greeness level
to_plot <- dfresult %>%
dplyr::left_join(temperatureToPlot, by=c("ordering", "status", "Rating")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private"))) %>%
{.}
to_plot_least <- to_plot %>%
dplyr::filter(`Rating`=="Least Green") %>%
{.}
to_plot_most <- to_plot %>%
dplyr::filter(`Rating`=="Most Green") %>%
{.}
binsize = 4
jitterAmount = 0.15
xlabel = "Temperature Bin"
## facet on public private and greeness level
to_plot %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value, color=Rating)) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable, color=Rating)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
## ggplot2::geom_segment(ggplot2::aes(x=`ordering`, xend=`ordering`, y=`count_start`, yend=`count_end`), size=7) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`-jitterAmount, xend=`ordering`-jitterAmount, y=`count_start`, yend=`count_end`, group=`Rating`, color=`Rating`), data=to_plot_least, size=binsize) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`+jitterAmount, xend=`ordering`+jitterAmount, y=`count_start`, yend=`count_end`, group=`status`, color=`Rating`), data=to_plot_most, size=binsize) +
## for slide
ggplot2::facet_wrap(.~`status`, ncol=2) +
## for paper
## ggplot2::facet_wrap(`Rating`~`status`, ncol=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(sprintf("ln(%s in kBtu / gross sqft) * 1000", ylabel)) +
## ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::theme_bw() +
ggplot2::labs(title=captions[i],
subtitle=(sprintf("%s model (greeness source: %s)", modelsuffix, green.source))) +
ggplot2::scale_color_manual(values=c("#A1D99B", "#31A354")) +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/facet_pubpri_bin_coef_ci_greeness%s_separate_organization_%s_slide%s%s.png", filenamesuffix, modelsuffix, energy_type, s), width=8, height=5, units="in")
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness_separate_organization_greeness_%s.png", modelsuffix), width=6, height=7, units="in")
}
allDataGreen %>%
distinct(Organization, Greeness, Name) %>%
dplyr::group_by(Organization, Greeness) %>%
dplyr::summarise(n())
## produce coefficient plot by climate region for separate portfolio using connected dashed lines for conf int
breaklevels = c("avgTempBelow10", "avgTemp10to20", "avgTemp20to30", "avgTemp30to40",
"avgTemp40to50", "avgTemp50to60", "avgTemp60to70", "avgTemp70to80", "avgTemp80to90",
"avgTempAbove90")
breaklabels = c("<10", "[10-20)", "[20-30)", "[30-40)",
"[40-50)", "[50-60)", "[60-70)", "[70-80)", "[80-90)",
">90")
modelname = "baseline"
xlabel = "Temperature Bin"
process_climate_result <- function(modelname, modellabel, energy_type, region_keyword, folder_suffix) {
## produce the bin model coefficient and CI plot for greeness analysis for the model for just one portfolio
df_recode_xi =
readr::read_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/header_to_%s_owner.csv", region_keyword)) %>%
{.}
filename = sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/reg_result%s/regression_%s_%s%s.txt",
folder_suffix, region_keyword, modelname, energy_type)
print(filename)
fields_to_remove = c("0b.", "eDate", "eState", "hasretrofit", "Constant", "o.avgTempBelow10")
fields_to_keep = NULL
outreg_vars = c("coef", "se", "ci_low", "ci_high")
vars_to_keep = c("coef", "ci_low", "ci_high")
df_model = read_outreg2(filename=filename, fields_to_remove=fields_to_remove, fields_to_keep=fields_to_keep,
outreg_vars=outreg_vars, vars_to_keep=vars_to_keep)
## df_model =
## readr::read_tsv(),
## skip=7, col_names=FALSE) %>%
## ## remove bottom non-conformative rows
## head(-5) %>%
## dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
## tidyr::fill(`X1`) %>%
## tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## ## modify this to X2:X?, ? is the number of models + 1
## tidyr::gather(`header`, `value`, X2:X9) %>%
## tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
## dplyr::filter(`variable` != "se",
## !grepl("0b.", `X1`, fixed = TRUE),
## !grepl("eDate", `X1`, fixed = TRUE),
## !grepl("eState", `X1`, fixed = TRUE),
## !grepl("hasretrofit", `X1`, fixed = TRUE),
## !grepl("Constant", `X1`, fixed = TRUE),
## !grepl("o.avgTempBelow10", `X1`, fixed = TRUE)
## ) %>%
df_zero = data.frame(X1="avgTemp60to70", variable=rep(c("coef", "ci_low", "ci_high"), 8),
header=sprintf("X%s", sort(rep(2:(length(unique(df_recode_xi$header)) + 1), 3))), value=0)
print(df_zero, n=Inf)
df_model <- df_model %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
tidyr::gather(`header`, `value`, starts_with("X")) %>%
## tidyr::gather(`header`, `value`, X2:X9) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
dplyr::bind_rows(df_zero) %>%
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::left_join(df_recode_xi, by="header") %>%
dplyr::mutate(`model`=modellabel) %>%
{.}
return(df_model)
}
breaklevels = c("avgTempBelow10", "avgTemp10to20", "avgTemp20to30", "avgTemp30to40",
"avgTemp40to50", "avgTemp50to60", "avgTemp60to70", "avgTemp70to80", "avgTemp80to90",
"avgTempAbove90")
breaklabels = c("<10", "[10-20)", "[20-30)", "[30-40)",
"[40-50)", "[50-60)", "[60-70)", "[70-80)", "[80-90)",
">90")
## energy_type = "_gas" ## use "" for electricity
energy_type = "_total"
## region_keyword = "climateregion"
## region_keyword = "usregion"
region_keyword = "region_wiki"
## regioncol = "US Climate Region"
## regioncol = "USRegion"
regioncol = "region_wiki"
## modelname = "statetrend"
## modelname = "baseline"
modelname = "both"
folder.suffixes = c("",
"_o_o_owned", "_o_oct_owned", "_r_o_owned", "_ro_oct_owned",
"_o_o_owned_and_leased", "_o_oct_owned_and_leased", "_r_o_owned_and_leased", "_ro_oct_owned_and_leased"
)
captions = c("no restrictions",
"Public: owned offices; Private: owned offices",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices",
"Public: owned offices; Private: owned retails",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices, retails",
"Public: offices; Private: offices",
"Public: offices, courthouses, CT/Offices; Private: offices",
"Public: offices; Private: retails",
"Public: offices, courthouses, CT/Offices; Private: offices, retails")
for (i in seq_along(folder.suffixes)) {
i = 1
s = folder.suffixes[i]
print(s)
tokens = unlist(strsplit(s, "_"))
if (length(tokens) == 0) {
ownlease = NA
usetype = NA
} else {
usetype = paste0(tokens[2:3], collapse = "_")
if ("leased" %in% tokens) {
ownlease = NA
} else {
ownlease = "Owned"
}
}
dfresult =
process_climate_result(modelname="baseline", modellabel="baseline", energy_type=energy_type,
region_keyword=region_keyword, s) %>%
dplyr::bind_rows(process_climate_result(modelname="statetrend", modellabel="state specific trend",
energy_type=energy_type,
region_keyword=region_keyword, s)) %>%
{.}
## remove 10F and 90F bins
dfresult <- dfresult %>%
dplyr::filter(!(`X1` %in% c("avgTempBelow10", "avgTempAbove90"))) %>%
{.}
## head(dfresult)
lower = min(dfresult$value)
upper = max(dfresult$value)
## (dfresult) %>%
## distinct(!!rlang::sym(regioncol), `status`, `model`)
## if (energy_type == "") {
## shifty=-400
## yupperlimit = 750
## scaling = 1 / 1e3 * 1.5
## } else if (energy_type == "_gas") {
## shifty=-900
## yupperlimit = 900
## scaling = 1 / 1e3 * 1.5
## } else if (energy_type == "_total") {
## ## shifty=-130
## ## yupperlimit = 220
## ## scaling = 1 / 1e4 * 4
## shifty=-50
## yupperlimit = 80
## scaling = 1 / 1e4 * 1.5
## }
temperatureBinData = get.study.set(ownlease, usetype)$df
temperatureSummary <- temperatureBinData %>%
dplyr::select(`Organization`, !!rlang::sym(regioncol), `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
dplyr::group_by(`Organization`, !!rlang::sym(regioncol), `tempBin`) %>%
dplyr::summarise(count = sum(value)) %>%
dplyr::ungroup() %>%
{.}
shifty = (-1)*(upper - lower)/2
scaling = abs(shifty) / max(temperatureSummary$count) / 2
temperatureToPlot <- temperatureSummary %>%
dplyr::mutate(`count`=`count` * scaling) %>%
dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
dplyr::select(`count_start`, `count_end`, `ordering`, `status`, !!rlang::sym(regioncol)) %>%
{.}
head(temperatureToPlot)
to_plot <- dfresult %>%
dplyr::left_join(temperatureToPlot, by=c("ordering", "status", regioncol)) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private"))) %>%
{.}
to_plot_public <- to_plot %>%
dplyr::filter(`status`=="public") %>%
{.}
to_plot_private <- to_plot %>%
dplyr::filter(`status`=="private") %>%
{.}
to_plot_north <- to_plot %>%
dplyr::filter(!!rlang::sym(regioncol)=="North") %>%
{.}
to_plot_south <- to_plot %>%
dplyr::filter(!!rlang::sym(regioncol)=="South") %>%
{.}
## for paper
## binsize = 4
## jitterAmount = 0.3
## for slides
binsize = 1.7
jitterAmount = 0.15
if (energy_type == "_total") {
binsize = 3.4
jitterAmount = 0.17
}
## facet by portfolio
p <- to_plot %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value, color=!!rlang::sym(regioncol)), size=0.7) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable, color=!!rlang::sym(regioncol))) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`-jitterAmount, xend=`ordering`-jitterAmount, y=`count_start`, yend=`count_end`, group=!!rlang::sym(regioncol), color=!!rlang::sym(regioncol)), data=to_plot_north, size=binsize) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`+jitterAmount, xend=`ordering`+jitterAmount, y=`count_start`, yend=`count_end`, group=!!rlang::sym(regioncol), color=!!rlang::sym(regioncol)), data=to_plot_south, size=binsize) +
## for paper
## ggplot2::facet_wrap(`US Climate Region`~`model`, ncol=2) +
## for slides
## use nrow=2 for climate region plots
ggplot2::facet_wrap(as.formula(paste("model~", "status")), ncol=2) +
## ggplot2::facet_wrap(`model`~`USRegion`, nrow=2) +
ggplot2::xlab("Temperature Bin") +
ggplot2::ylab("ln(Electricity in kBtu / gross sqft) * 1000") +
## ggplot2::coord_cartesian(ylim=c(shifty, yupperlimit)) +
## this is to plot slide_zoom for electric
## ggplot2::ylim(c(-100, 100)) +
## this is to plot slide_zoom for gas
## ggplot2::ylim(c(-150, 150)) +
ggplot2::theme_bw() +
ggplot2::scale_color_manual(values = RColorBrewer::brewer.pal(3, "RdBu")[c(3, 1)], name="US Region") +
ggplot2::ggtitle(captions[i]) +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
if (regioncol == "US Climate Region") {
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate_slide_zoom%s.png", energy_type), width=8, height=5, units="in")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate_slide%s.png", energy_type), width=8, height=5, units="in")
## for paper
## ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate.png", width=8, height=7, units="in")
}
else {
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_%s_slide_zoom%s.png", regioncol, energy_type), width=8, height=5, units="in")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_%s_slide%s_%s%s.png", regioncol, energy_type, modelname, s), width=8, height=6, units="in")
## for paper
## ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_usregion.png", width=8, height=7, units="in")
}
## facet by region
p <- to_plot %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value, color=status), size=0.7) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable, color=status)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`-jitterAmount, xend=`ordering`-jitterAmount, y=`count_start`, yend=`count_end`, group=!!rlang::sym("status"), color=!!rlang::sym("status")), data=to_plot_public, size=binsize) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`+jitterAmount, xend=`ordering`+jitterAmount, y=`count_start`, yend=`count_end`, group=!!rlang::sym("status"), color=!!rlang::sym("status")), data=to_plot_private, size=binsize) +
## for paper
## ggplot2::facet_wrap(`US Climate Region`~`model`, ncol=2) +
## for slides
## use nrow=2 for climate region plots
ggplot2::facet_wrap(as.formula(paste("model~", regioncol)), ncol=2) +
## ggplot2::facet_wrap(`model`~`USRegion`, nrow=2) +
ggplot2::xlab("Temperature Bin") +
ggplot2::ylab("ln(Electricity in kBtu / gross sqft) * 1000") +
## ggplot2::coord_cartesian(ylim=c(shifty, yupperlimit)) +
## this is to plot slide_zoom for electric
## ggplot2::ylim(c(-100, 100)) +
## this is to plot slide_zoom for gas
## ggplot2::ylim(c(-150, 150)) +
ggplot2::ggtitle(captions[i]) +
ggplot2::theme_bw() +
ggplot2::scale_color_grey() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
print(p)
if (regioncol == "US Climate Region") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate_slide_zoom%s_facetRegion.png", energy_type), width=8, height=5, units="in")
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate_slide%s_facetRegion.png", energy_type), width=8, height=5, units="in")
## for paper
## ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate_facetRegion.png", width=8, height=7, units="in")
}
else {
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_%s_slide_zoom%s_facetRegion.png", regioncol, energy_type), width=8, height=5, units="in")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_%s_slide%s_%s_facetRegion%s.png", regioncol, energy_type, modelname, s), width=8, height=5, units="in")
## for paper
## ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_usregion_facetRegion.png", width=8, height=7, units="in")
}
}
## generate image tex file
newlines = NULL
## imagePrefix = "bin_coef_ci_region_wiki_slide_total_both_facetRegion"
## filename = "facetRegionPlots"
## imagePrefix = "bin_coef_ci_region_wiki_slide_total_both"
## filename = "facetPortfolioPlots"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_separate_organization_baseline_slide_total"
## filename = "facetGreenessPlots_baseline_majority"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_howestate_separate_organization_baseline_slide_total"
## filename = "facetGreenessPlots_baseline_howestate"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_howecounty_separate_organization_baseline_slide_total"
## filename = "facetGreenessPlots_baseline_howecounty"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_separate_organization_statetrend_slide_total"
## filename = "facetGreenessPlots_statetrend_majority"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_howestate_separate_organization_statetrend_slide_total"
## filename = "facetGreenessPlots_statetrend_howestate"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_howecounty_separate_organization_statetrend_slide_total"
## filename = "facetGreenessPlots_statetrend_howecounty"
## imagePrefix = "bin_coef_ci_separate_organization_baseline_slide_total"
## filename = "main_analysis_baseline_total"
imagePrefix = "bin_coef_ci_separate_organization_baseline_slide_gas"
filename = "main_analysis_baseline_gas"
## imagePrefix = "bin_coef_ci_separate_organization_baseline_slide"
## filename = "main_analysis_baseline"
for (s in folder.suffixes) {
newlines <- c(newlines,
"\\begin{frame}",
"\\begin{figure}[h!]",
"\\centering",
sprintf("\\includegraphics[width=1.0\\linewidth]{images/%s%s.png}",
imagePrefix, s),
"\\end{figure}",
"\\end{frame}")
}
con <- file(sprintf("~/Dropbox/thesis/writeups/policy_cmp/images/%s.tex", filename), open = "w+")
writeLines(newlines, con, sep = "\n", useBytes = FALSE)
close(con)
## generate tex order by building type first then by greenness source
for (modeltype in c("baseline", "statetrend")) {
newlines = NULL
filename = sprintf("facetGreenessPlots_%s", modeltype)
for (s in folder.suffixes) {
for (green.suffix in c("", "_howestate", "_howecounty")) {
imagePrefix = sprintf("facet_pubpri_bin_coef_ci_greeness%s_separate_organization_%s_slide_total",
green.suffix, modeltype)
newlines <- c(newlines,
"\\begin{frame}",
"\\begin{figure}[h!]",
"\\centering",
sprintf("\\includegraphics[width=1.0\\linewidth]{images/%s%s.png}",
imagePrefix, s),
"\\end{figure}",
"\\end{frame}")
}
con <- file(sprintf("~/Dropbox/thesis/writeups/policy_cmp/images/%s.tex", filename), open = "w+")
writeLines(newlines, con, sep = "\n", useBytes = FALSE)
close(con)
}
}
## generate regression result summary table for bin model with greeness
## portfolio="PNC"
## status_code = "private"
portfolio="GSA"
status_code = "public"
summary_portfolio = readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_bin_greeness_0720_%s.txt", portfolio), skip=7, col_names=FALSE) %>%
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
dplyr::filter(!(`variable` %in% c("ci_low", "ci_high")),
!grepl("1b.eGreeness", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE)
) %>%
dplyr::mutate(`X1`=gsub("2.eGreeness#c.", "Moderate Green x ", `X1`)) %>%
dplyr::mutate(`X1`=gsub("3.eGreeness#c.", "Most Green x ", `X1`)) %>%
dplyr::mutate(`X1`=ifelse(as.numeric(rownames(.)) %% 2 == 1, `X1`, "")) %>%
{.}
tail_summary_portfolio = readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_bin_greeness_0720_%s.txt", portfolio), skip=7, col_names=FALSE) %>%
tail(5) %>%
{.}
summary_export = summary_portfolio %>%
dplyr::bind_rows(tail_summary_portfolio) %>%
{.}
summary_export %>%
readr::write_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_greeness_model_%s.csv", portfolio))
summary_export <-
summary_export %>%
dplyr::select(-`variable`) %>%
dplyr::rename(" "=`X1`, "(1)"=`X2`, "(2)"=`X3`) %>%
{.}
result_table_tex <- xtable(summary_export, caption=sprintf("Regression result for the %s portfolio", status_code))
print(result_table_tex, tabular.environment = "longtable", include.rownames=FALSE,
file=sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_greeness_result.tex", portfolio))
# long table of state temperature
avg_temp_state =
readr::read_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/stateAverageTemperature2011to2013.csv") %>%
tibble::as_data_frame() %>%
dplyr::select(`Location`, `Value`) %>%
dplyr::rename(`Average temperature (F)`=`Value`, `State`=`Location`) %>%
{.}
result_table_tex <- xtable(avg_temp_state, caption="Average state temperature from 2011 to 2013")
print(result_table_tex, tabular.environment = "longtable", include.rownames=FALSE,
file="~/Dropbox/thesis/writeups/policy_cmp/tables/avg_state_temp.tex", size="\\footnotesize")
## only plot the temperature bin model
models = "Temperature Bin"
plotHeight = 6
## plot all models
## models=c("No Control", "HDD CDD", "Polynomial degree 2 HDD and CDD", "Temperature Bin")
## warmth_class_method = "building temperature"
## warmth_class_method = "state latitude"
warmth_class_method = "state temperature"
## plot coefficients for retrofit climate event study
pd <- position_dodge(width = 0.4)
## quarterly plot setting starts -------------------------
duration = "quarterly"
energy_type = ""
## energy_type = "_gas"
endbound = 20
fatten_amount = 2.5
size = 0.2
point_size = 1.0
## quarterly plot setting ends -------------------------
## monthly plot setting starts -------------------------
## duration = "monthly"
## energy_type = "_gas"
## endbound = 60
## fatten_amount = 2.5
## size = 0.2
## point_size = 1.0
## monthly plot setting ends -------------------------
## yearly plot setting starts -------------------------
## duration = "yearly"
## endbound = 5
## fatten_amount = 2.5
## size = 1.0
## point_size = 2.5
## yearly plot setting ends -------------------------
retrofits = c("Advanced Metering", "Building Envelope", "HVAC", "Lighting", "GSALink", "Bundle")
names(retrofits) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
if (energy_type == "") {
energy_label = "Electricity"
} else if (energy_type == "_gas") {
energy_label = "Gas"
}
## for (fileSuffix in c("metering")) {
for (fileSuffix in c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")) {
retrofit = retrofits[[fileSuffix]]
if (warmth_class_method == "building temperature") {
p <-
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_buildingTemp_event_study_%s%s.txt", duration, fileSuffix, energy_type), skip=7, col_names=FALSE)
} else if (warmth_class_method == "state temperature") {
p <-
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_event_study_%s%s.txt", duration, fileSuffix, energy_type), skip=7, col_names=FALSE)
} else if (warmth_class_method == "state latitude") {
p <-
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_stateLat_event_study_%s%s.txt", duration, fileSuffix, energy_type), skip=7, col_names=FALSE)
}
## remove bottom non-conformative rows
p <- p %>%
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`model`, `value`, X2:X5) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("o.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("0b.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("3b.eWarmth", `X1`, fixed = TRUE),
!grepl("o.eWarmth", `X1`, fixed = TRUE),
!grepl("cdd", `X1`, fixed=TRUE),
!grepl("hdd", `X1`, fixed=TRUE),
!grepl("Constant", `X1`, fixed=TRUE),
!grepl("avgTemp", `X1`, fixed=TRUE),
) %>%
{.}
if (duration == "year") {
p <- p %>%
dplyr::filter(!grepl(".year", `X1`, fixed = TRUE)) %>%
{.}
} else if ((duration == "monthly") | (duration == "quarterly")){
print("asdf")
p <- p %>%
dplyr::filter(!grepl(".eDate", `X1`, fixed = TRUE)) %>%
{.}
}
p <- p %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=gsub(",", "", `value`)) %>%
tidyr::unite(`temp`, `X1`, `model`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "model"), sep="-") %>%
dplyr::filter(!is.na(`coef`)) %>%
## readr::write_csv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_coef_ci.csv", fileSuffix))
dplyr::mutate(`coef`=as.numeric(`coef`),
`ci_high`=as.numeric(`ci_high`),
`ci_low`=as.numeric(`ci_low`),
) %>%
dplyr::mutate(warmth = ifelse(grepl("1.eWarmth", X1, fixed=TRUE), "cold",
ifelse(grepl("2.eWarmth", X1, fixed=TRUE), "hot", "mild"))) %>%
dplyr::mutate(`tau`=as.integer(substr(X1, 1, stringi::stri_locate(X1, fixed=".", mode="first")-1))) %>%
dplyr::mutate(`tau`=ifelse(`tau` <= endbound, `tau`, `tau` - 2*endbound - 1)) %>%
dplyr::mutate(`warmth`=ifelse(as.numeric(tau) < 0, "pre-retrofit all warmth level", `warmth`)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="No Control", "X3"="HDD CDD", "X4"="Polynomial degree 2 HDD and CDD",
"X5"="Temperature Bin") %>%
dplyr::select(-`X1`) %>%
dplyr::bind_rows(data.frame(ci_low=NA, ci_high=NA, coef=0, term=NA,
warmth=rep(c("cold", "hot", "mild", "pre-retrofit all warmth level"), 4),
model=c(rep("No Control", 4), rep("HDD CDD", 4),
rep("Polynomial degree 2 HDD and CDD", 4), rep("Temperature Bin", 4)),
tau=0)) %>%
{.}
print(head(p))
p <- p %>%
dplyr::filter(`model` %in% models) %>%
{.}
p <- p %>%
ggplot2::ggplot(ggplot2::aes(x=tau, y=coef, color=warmth, group=warmth)) +
ggplot2::geom_line() +
## ggplot2::geom_point() +
ggplot2::geom_pointrange(ggplot2::aes(x=tau, y=`coef`, ymin=`ci_low`, ymax=`ci_high`, color=warmth, group=warmth), position=pd, shape=18, fatten=fatten_amount, size=size) +
ggplot2::geom_point(x=0, y=0, shape=18, size=point_size) +
ggplot2::facet_wrap(`model`~., ncol=2) +
ggplot2::xlab("Number of years before of after retrofit") +
ggplot2::ylab(NULL) +
## ggplot2::coord_cartesian(ylim=c(-20, 25)) +
## ggplot2::scale_color_brewer(palette="Spectral") +
## ggplot2::scale_color_brewer(palette="RdBu") +
## ggplot2::scale_color_manual(values=c("#5698C4", "#E97550", "grey")) +
## ggplot2::scale_color_manual(values=c("#d55e00", "#8B8CB0", "skyblue")) +
## ggplot2::scale_color_manual(values=c("#d55e00", "#009e73", "skyblue")) +
## logan's
## ggplot2::scale_color_manual(values=colorRampPalette(c("#FF0000","#000066"))(3L)) +
ggplot2::scale_color_manual(values=c("#67A9CF", "#EF8A62", "grey70", "black")) +
## ggplot2::theme_dark() +
ggplot2::theme_bw() +
ggtitle(sprintf("%s Event study \"%s\" retrofit\nWarmth level by %s",
energy_label, retrofit, warmth_class_method)) +
ggplot2::theme(legend.position = "bottom",
plot.title = element_text(size=slide_size))
print(p)
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_retrofit_coef_noci_%s.png", duration, fileSuffix), width=8, height=5, units="in")
if (warmth_class_method == "building temperature") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_buildingTemp_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type), width=8, height=plotHeight, units="in")
} else if (warmth_class_method == "state temperature") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type), width=8, height=plotHeight, units="in")
} else if (warmth_class_method == "state latitude") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_stateLat_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type), width=8, height=plotHeight, units="in")
}
}
df_recode_retrofit =
readr::read_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/separate_model_monthly_header_to_retrofit.csv") %>%
{.}
## plot coefficients for retrofit climate event study with each warmth level fitted separately
## plot all models
## models = c("no_control", "hddcdd", "hdd2cdd2", "bin")
## plotHeight = 5
## plot one model
models = "bin"
plotHeight = 6
slide_size = 22
## warmth_class_method = "state temperature"
## warmth_class_method = "state latitude"
warmth_class_method = "building temperature"
pd <- position_dodge(width = 0.4)
## quarterly plot setting starts -------------------------
duration = "quarterly"
## energy_type = ""
energy_type = "_gas"
endbound = 20
fatten_amount = 2.5
size = 0.2
point_size = 1.0
## quarterly plot setting ends -------------------------
## monthly plot setting starts -------------------------
## duration = "monthly"
## energy_type = "_gas"
## endbound = 60
## fatten_amount = 2.5
## size = 0.2
## point_size = 1.0
## monthly plot setting ends -------------------------
## yearly plot setting starts -------------------------
## duration = "yearly"
## endbound = 5
## fatten_amount = 2.5
## size = 1.0
## point_size = 2.5
## yearly plot setting ends -------------------------
if (energy_type == "") {
energy_label = "Electricity"
} else if (energy_type == "_gas") {
energy_label = "Gas"
}
acc <- lapply(models, function(model_type) {
if (warmth_class_method == "state temperature") {
dfacc =
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/separate_model_%s_retrofit_climate_%s%s.txt", duration, model_type, energy_type), skip=7, col_names=FALSE)
} else if (warmth_class_method == "state latitude") {
dfacc =
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/separate_model_%s_stateLat_retrofit_climate_%s%s.txt", duration, model_type, energy_type), skip=7, col_names=FALSE)
} else if (warmth_class_method == "building temperature") {
dfacc =
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/separate_model_%s_buildingTemp_retrofit_climate_%s%s.txt", duration, model_type, energy_type), skip=7, col_names=FALSE)
}
dfacc <- dfacc %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`header`, `value`, X2:X19) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl(".eDate", `X1`, fixed = TRUE),
!grepl("o.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("0b.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("cdd", `X1`, fixed=TRUE),
!grepl("hdd", `X1`, fixed=TRUE),
!grepl("Constant", `X1`, fixed=TRUE),
!grepl("avgTemp", `X1`, fixed=TRUE),
) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=gsub(",", "", `value`)) %>%
tidyr::unite(`temp`, `X1`, `header`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "header"), sep="-") %>%
dplyr::filter(!is.na(`coef`)) %>%
dplyr::mutate(`coef`=as.numeric(`coef`),
`ci_high`=as.numeric(`ci_high`),
`ci_low`=as.numeric(`ci_low`),
) %>%
dplyr::mutate(`tau`=as.integer(substr(X1, 1, stringi::stri_locate(X1, fixed=".", mode="first")-1))) %>%
dplyr::mutate(`tau`=ifelse(`tau` <= endbound, `tau`, `tau` - 2*endbound - 1)) %>%
dplyr::left_join(df_recode_retrofit, by="header") %>%
dplyr::mutate(`model`=model_type) %>%
dplyr::select(-`X1`, -`header`) %>%
{.}
})
modelResult = do.call(rbind, acc)
retrofits = c("Advanced Metering", "Building Envelope", "HVAC", "Lighting", "GSALink", "Bundle")
names(retrofits) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
for (fileSuffix in names(retrofits)) {
retrofit_type <- retrofits[[fileSuffix]]
print(retrofit_type)
p <- modelResult %>%
dplyr::bind_rows(data.frame(`tau`=0, `ci_low`=0, `ci_high`=0, `coef`=0, `retrofit`=retrofit_type,
`warmth`=rep(c("mild", "cold", "hot"), 4),
`model`=c(rep("no_control", 3), rep("hddcdd", 3), rep("hdd2cdd2", 3),
rep("bin", 3)))) %>%
dplyr::filter(`model` %in% models) %>%
dplyr::filter(retrofit == retrofit_type) %>%
ggplot2::ggplot(ggplot2::aes(x=tau, y=coef, color=warmth, group=warmth)) +
ggplot2::geom_line() +
ggplot2::geom_pointrange(ggplot2::aes(x=tau, y=`coef`, ymin=`ci_low`, ymax=`ci_high`, color=warmth, group=warmth), position=pd, shape=18, fatten=fatten_amount, size=size) +
ggplot2::geom_point(x=0, y=0, shape=18, size=point_size) +
ggplot2::facet_wrap(`model`~., ncol=2) +
ggplot2::xlab("Number of years before of after retrofit") +
ggplot2::ylab(NULL) +
ggplot2::scale_color_manual(values=c("#67A9CF", "#EF8A62", "grey70", "black")) +
ggplot2::theme_bw() +
ggtitle(sprintf("%s Event study \"%s\" retrofit\nWarmth level by %s", energy_label, retrofit_type, warmth_class_method)) +
ggplot2::theme(legend.position = "bottom",
plot.title = element_text(size=slide_size))
print(p)
if (warmth_class_method == "state temperature") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/sep_%s_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type),width=8, height=plotHeight, units="in")
} else if (warmth_class_method == "state latitude") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/sep_%s_stateLat_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type), width=8, height=plotHeight, units="in")
} else if (warmth_class_method == "building temperature") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/sep_%s_buildingTemp_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type), width=8, height=plotHeight, units="in")
}
}
## generate summary statistics tables
retrofits = c("Advanced Metering", "Building Envelope", "HVAC", "Lighting", "GSALink", "")
names(retrofits) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
## for (fileSuffix in c("metering")) {
for (fileSuffix in c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")) {
retrofit = retrofits[[fileSuffix]]
summary_head = readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/yearly_event_study_%s.txt", fileSuffix), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
dplyr::filter(!(`variable` %in% c("ci_low", "ci_high"))) %>%
dplyr::filter(
## `variable` != "se",
!grepl(".year", `X1`, fixed = TRUE),
!grepl("o.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("0b.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("3b.eWarmth", `X1`, fixed = TRUE),
!grepl("o.eWarmth", `X1`, fixed = TRUE),
) %>%
dplyr::mutate(X1 = gsub("1.eWarmth", "cold", X1)) %>%
dplyr::mutate(X1 = gsub("2.eWarmth", "hot", X1)) %>%
dplyr::mutate(X1 = gsub("1.tau_bounded_post", "$I[\\\\tau=1]$", X1)) %>%
dplyr::mutate(X1 = gsub("2.tau_bounded_post", "$I[\\\\tau=2]$", X1)) %>%
dplyr::mutate(X1 = gsub("3.tau_bounded_post", "$I[\\\\tau=3]$", X1)) %>%
dplyr::mutate(X1 = gsub("4.tau_bounded_post", "$I[\\\\tau=4]$", X1)) %>%
dplyr::mutate(X1 = gsub("5.tau_bounded_post", "$I[\\\\tau>=5]$", X1)) %>%
dplyr::mutate(X1 = gsub("1.tau_bounded_shift", "$I[\\\\tau=1]$", X1)) %>%
dplyr::mutate(X1 = gsub("2.tau_bounded_shift", "$I[\\\\tau=2]$", X1)) %>%
dplyr::mutate(X1 = gsub("3.tau_bounded_shift", "$I[\\\\tau=3]$", X1)) %>%
dplyr::mutate(X1 = gsub("4.tau_bounded_shift", "$I[\\\\tau=4]$", X1)) %>%
dplyr::mutate(X1 = gsub("5.tau_bounded_shift", "$I[\\\\tau=5]$", X1)) %>%
dplyr::mutate(X1 = gsub("6.tau_bounded_shift", "$I[\\\\tau<=-5]$", X1)) %>%
dplyr::mutate(X1 = gsub("7.tau_bounded_shift", "$I[\\\\tau=-4]$", X1)) %>%
dplyr::mutate(X1 = gsub("8.tau_bounded_shift", "$I[\\\\tau=-3]$", X1)) %>%
dplyr::mutate(X1 = gsub("9.tau_bounded_shift", "$I[\\\\tau=-2]$", X1)) %>%
dplyr::mutate(X1 = gsub("10.tau_bounded_shift", "$[\\\\tau=-1]$", X1)) %>%
dplyr::mutate(X1 = gsub("#", "$\\\\times$", X1)) %>%
dplyr::mutate(`X1`=ifelse(as.numeric(rownames(.)) %% 2 == 1, `X1`, "")) %>%
## dplyr::select(-`variable`) %>%
{.}
print(head(summary_head))
summary_tail = readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/yearly_event_study_%s.txt", fileSuffix), skip=7, col_names=FALSE) %>%
tail(5) %>%
dplyr::mutate(`X1` = gsub(" ci_high", "", `X1`)) %>%
dplyr::mutate(`X1` = gsub("<", "$<$", `X1`)) %>%
{.}
print(head(summary_tail))
summary_export = summary_head %>%
dplyr::bind_rows(summary_tail) %>%
{.}
summary_export %>%
readr::write_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_retrofit_model_%s.csv", fileSuffix))
summary_export <-
summary_export %>%
dplyr::select(-`variable`) %>%
dplyr::rename(" "=`X1`, "(1)"=`X2`, "(2)"=`X3`, "(3)"=`X4`, "(4)"=`X5`) %>%
{.}
result_table_tex <- xtable(summary_export, caption=sprintf("Regression result for event study model of buildings with %s retrofit", fileSuffix))
print(result_table_tex, tabular.environment = "longtable", include.rownames=FALSE,
file=sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/summary_retrofit_model_%s.tex", fileSuffix), size="\\footnotesize", sanitize.text.function=function(x){x})
## above sanitize function overwritten so that the math symbols stay the original way
}
## plot coefficients for retrofit climate event study with each warmth level fitted separately
pd <- ggplot2::position_dodge(width = 0.4)
## quarterly plot setting starts -------------------------
energy_type = ""
## energy_type = "_gas"
duration = "quarterly"
additive_keywords = "nonsingle_"
if (additive_keywords == "") {
df_recode_retrofit = data.frame(header=sprintf("X%s", 2:7),
retrofit=c("Advanced Metering", "Bundle", "Building Envelope", "GSALink", "HVAC",
"Lighting"))
} else if (additive_keywords == "nonsingle_") {
df_recode_retrofit = data.frame(header=sprintf("X%s", 2:7),
retrofit=c("Advanced Metering", "Building Envelope", "Commissioning", "GSALink",
"HVAC", "Lighting"))
}
endbound = 20
fatten_amount = 2.5
size = 0.2
point_size = 1.0
## quarterly plot setting starts -------------------------
## monthly plot setting starts -------------------------
## energy_type = ""
## ## energy_type = "_gas"
## duration = "monthly"
## endbound = 60
## fatten_amount = 2.5
## size = 0.2
## point_size = 1.0
## produce plots with all 4 model type
## models = c("no_control", "hddcdd", "hdd2cdd2", "bin")
## produce plots with only one model type
models = c("bin")
## monthly plot setting ends -------------------------
## yearly plot setting starts -------------------------
## duration = "yearly"
## endbound = 5
## fatten_amount = 2.5
## size = 1.0
## point_size = 2.5
## yearly plot setting ends -------------------------
acc <- lapply(models, function(model_type) {
dfacc =
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_%sretrofit_no_climate_%s%s.txt", duration, additive_keywords, model_type, energy_type), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`header`, `value`, X2:X7) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl(".eDate", `X1`, fixed = TRUE),
!grepl("cdd", `X1`, fixed=TRUE),
!grepl("hdd", `X1`, fixed=TRUE),
!grepl("Constant", `X1`, fixed=TRUE),
!grepl("avgTemp", `X1`, fixed=TRUE),
) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=gsub(",", "", `value`)) %>%
tidyr::unite(`temp`, `X1`, `header`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "header"), sep="-") %>%
dplyr::filter(!is.na(`coef`)) %>%
dplyr::mutate(`coef`=as.numeric(`coef`),
`ci_high`=as.numeric(`ci_high`),
`ci_low`=as.numeric(`ci_low`),
) %>%
dplyr::mutate(`tau`=as.integer(substr(X1, 1, stringi::stri_locate(X1, fixed=".", mode="first")-1))) %>%
dplyr::mutate(`tau`=ifelse(`tau` <= endbound, `tau`, `tau` - 2*endbound - 1)) %>%
dplyr::left_join(df_recode_retrofit, by="header") %>%
dplyr::mutate(`model`=model_type) %>%
dplyr::select(-`X1`, -`header`) %>%
{.}
})
modelResult = do.call(rbind, acc)
unique(modelResult$model)
tail(modelResult)
if (additive_keywords == "") {
retrofits = c("Advanced Metering", "Building Envelope", "HVAC", "Lighting", "GSALink", "Bundle")
names(retrofits) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
ylower = c(-1, -1, -4.5, -1.5, -5, -1)
names(ylower) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
yupper = c(1.2, 0.9, 2.2, 1.4, 5, 0.6)
names(yupper) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
} else if (additive_keywords == "nonsingle_") {
retrofits = c("Advanced Metering", "Building Envelope", "HVAC", "Lighting", "GSALink", "Commissioning")
names(retrofits) = c("metering", "envelope", "hvac", "lighting", "GSALink", "commissioning")
ylower = c(-0.6, -0.8, -0.6, -0.75, -2, -0.75)
names(ylower) = c("metering", "envelope", "hvac", "lighting", "GSALink", "commissioning")
yupper = c(0.2, 0.25, 0.2, 0.25, 1.5, 0.25)
names(yupper) = c("metering", "envelope", "hvac", "lighting", "GSALink", "commissioning")
}
if (energy_type == "") {
energy_label = "Electricity"
} else if (energy_type == "_gas") {
energy_label = "Gas"
}
slide_size = 22
for (fileSuffix in names(retrofits)) {
retrofit_type <- retrofits[[fileSuffix]]
print(retrofit_type)
p <-
modelResult %>%
dplyr::filter(retrofit == retrofit_type) %>%
dplyr::bind_rows(data.frame(`tau`=0, `ci_low`=0, `ci_high`=0, `coef`=0, `retrofit`=retrofit_type, `model`=models)) %>%
ggplot2::ggplot(ggplot2::aes(x=tau, y=coef)) +
ggplot2::geom_line() +
ggplot2::geom_pointrange(ggplot2::aes(x=tau, y=`coef`, ymin=`ci_low`, ymax=`ci_high`), position=pd, shape=18, fatten=fatten_amount, size=size) +
ggplot2::geom_point(x=0, y=0, shape=18, size=point_size)
if (length(models) == 1) {
p <- p +
ggplot2::facet_wrap(`model`~.)
} else {
p <- p +
ggplot2::facet_wrap(`model`~., ncol=2)
}
p <- p +
ggplot2::xlab("Number of periods before or after retrofit") +
ggplot2::ylab(NULL) +
ggplot2::theme_bw() +
ggtitle(sprintf("%s Event study model of \"%s\"", energy_label, retrofit_type))
if (!is.null(ylower)) {
p <- p +
ggplot2::coord_cartesian(ylim=c(ylower[[fileSuffix]], yupper[[fileSuffix]]))
}
p <- p +
ggplot2::theme(legend.position = "bottom",
plot.title = element_text(size=slide_size))
print(p)
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_%sretrofit_no_climate_coef_ci_%s_%s%s.png", duration, additive_keywords, fileSuffix, length(models), energy_type), width=8, height=5, units="in")
}
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
## start section: create summary of energy, weather, and building count for 4 alternative study set, based on different building types
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
get.study.set <- function(own.or.lease, office.class) {
if (!is.na(office.class)) {
if (office.class == "o_o") {
generic.office.pnc = "Office"
generic.office.gsa = "Office"
} else if (office.class == "o_oct") {
generic.office.pnc = "Office"
generic.office.gsa = c("Courthouse", "Office", "CT/Office")
} else if (office.class == "r_o") {
generic.office.pnc = "Retail"
generic.office.gsa = "Office"
} else if (office.class == "ro_oct") {
generic.office.pnc = c("Retail", "Office")
generic.office.gsa = c("Courthouse", "Office", "CT/Office")
}
to.summarize = allDataGreen %>%
dplyr::filter((Organization == "PNC" & `type_general` %in% generic.office.pnc) | (Organization == "GSA" & `type_general` %in% generic.office.gsa)) %>%
{.}
} else {
to.summarize = allDataGreen
generic.office.gsa = NA
generic.office.pnc = NA
}
if (!is.na(own.or.lease)) {
to.summarize <- to.summarize %>%
dplyr::filter(`Ownership`==own.or.lease) %>%
{.}
}
return(list(df=to.summarize, generic.office.gsa=generic.office.gsa, generic.office.pnc=generic.office.pnc))
}
## following is how to use the above function
## restrict to owned public offices, courthouse, private offices
## temp <- get.study.set("Owned", "o_oct")
## restrict to owned and leased public offices, courthouse, private offices
## temp <- get.study.set(NA, "o_oct")
## No restrictions
## temp <- get.study.set(NA, NA)
temp$df
## following check on what is returned by the above function
temp %>%
distinct(Name, `type_general`, Organization) %>%
dplyr::group_by(Organization, `type_general`) %>%
dplyr::summarise(n()) %>%
print()
folder.suffixes = c("",
"_o_o_owned", "_o_oct_owned", "_r_o_owned", "_ro_oct_owned",
"_o_o_owned_and_leased", "_o_oct_owned_and_leased", "_r_o_owned_and_leased", "_ro_oct_owned_and_leased"
)
captions = c("no restrictions",
"Public: owned offices; Private: owned offices",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices",
"Public: owned offices; Private: owned retails",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices, retails",
"Public: offices; Private: offices",
"Public: offices, courthouses, CT/Offices; Private: offices",
"Public: offices; Private: retails",
"Public: offices, courthouses, CT/Offices; Private: offices, retails")
## change the major source of greeness to howe by county
data.to.summarize = allDataGreen %>%
dplyr::mutate(Greeness = greeness.howe.county) %>%
{.}
for (i in seq_along(folder.suffixes)) {
s = folder.suffixes[i]
print(s)
tokens = unlist(strsplit(s, "_"))
if (length(tokens) == 0) {
ownlease = NA
usetype = NA
} else {
usetype = paste0(tokens[2:3], collapse = "_")
if ("leased" %in% tokens) {
ownlease = NA
} else {
ownlease = "Owned"
}
}
if (is.na(ownlease)) {
own.or.lease.str = ""
} else {
own.or.lease.str = ownlease
}
result <- get.study.set(ownlease, usetype)
to.summarize = result$df
generic.office.gsa = result$generic.office.gsa
generic.office.pnc = result$generic.office.pnc
## may not be the properway to check whether the vector is NA
if (is.na(generic.office.gsa[1])) {
generic.office.gsa = "no restriction"
}
if (is.na(generic.office.pnc[1])) {
generic.office.pnc = "no restriction"
}
to.summarize %>%
distinct(Name, `type_general`, Organization) %>%
dplyr::group_by(Organization, `type_general`) %>%
dplyr::summarise(n()) %>%
print()
df.n = to.summarize %>%
dplyr::distinct(Organization, Name) %>%
dplyr::group_by(Organization) %>%
dplyr::summarise(cnt=n()) %>%
dplyr::ungroup() %>%
tibble::column_to_rownames("Organization") %>%
{.}
## slide summary stats
df1 = to.summarize %>%
dplyr::mutate(`<40`=`<10` + `[10-20)` + `[20-30)` + `[30-40)`,
`>80`=`[80-90)` + `>90`) %>%
dplyr::select(`eui_elec`, `eui_gas`, `<40`, `>80`, `private`, `Name`) %>%
dplyr::group_by(`private`) %>%
dplyr::summarise(`eui_elec`=mean(`eui_elec`), `eui_gas`=mean(`eui_gas`), `<40`=mean(`<40`), `>80`=mean(`>80`), `building count`=n_distinct(Name)) %>%
dplyr::ungroup() %>%
{.}
df2 = to.summarize %>%
dplyr::mutate(`<40`=`<10` + `[10-20)` + `[20-30)` + `[30-40)`,
`>80`=`[80-90)` + `>90`) %>%
dplyr::select(`eui_elec`, `eui_gas`, `<40`, `>80`, `private`, `region_wiki`, `Name`) %>%
dplyr::group_by(`private`, `region_wiki`) %>%
dplyr::summarise(`eui_elec`=mean(`eui_elec`), `eui_gas`=mean(`eui_gas`), `<40`=mean(`<40`), `>80`=mean(`>80`), `building count`=n_distinct(Name)) %>%
dplyr::ungroup() %>%
{.}
df3 = to.summarize %>%
dplyr::mutate(`<40`=`<10` + `[10-20)` + `[20-30)` + `[30-40)`,
`>80`=`[80-90)` + `>90`) %>%
dplyr::select(`eui_elec`, `eui_gas`, `<40`, `>80`, `private`, `Greeness`, `Name`) %>%
dplyr::group_by(`private`, `Greeness`) %>%
dplyr::summarise(`eui_elec`=mean(`eui_elec`), `eui_gas`=mean(`eui_gas`), `<40`=mean(`<40`), `>80`=mean(`>80`), `building count`=n_distinct(Name)) %>%
dplyr::ungroup() %>%
{.}
sink(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/energy_weather_n_by_region_green%s.tex",
s))
df1 %>%
dplyr::bind_rows(df2, df3) %>%
dplyr::select(private, `region_wiki`, Greeness, everything()) %>%
dplyr::mutate(order = c(1, 6, 2, 3, 7, 8, 4, 5, 9, 10)) %>%
dplyr::arrange(order) %>%
## dplyr::mutate(private = c("public", rep("", 5), "private", rep("", 5))) %>%
dplyr::mutate(private=ifelse(private == 1, "private", "public")) %>%
dplyr::mutate(` `=ifelse(is.na(`region_wiki`), ifelse(is.na(`Greeness`), private, `Greeness`), `region_wiki`)) %>%
dplyr::select(-`region_wiki`, -`Greeness`, -private, -order) %>%
dplyr::select(` `, everything()) %>%
dplyr::rename(`Electricity`=`eui_elec`, `Gas`=`eui_gas`) %>%
knitr::kable("latex", booktabs = T, format.args=list(big.mark=",", digits=3),
caption=sprintf("%s PNC (n=%s): %s; GSA (n=%s): %s",
own.or.lease.str,
df.n["PNC",1],
paste(generic.office.pnc, collapse = ", "),
df.n["GSA",1],
paste(generic.office.gsa, collapse = ", "))) %>%
kableExtra::kable_styling(full_width = FALSE, font_size = 7) %>%
kableExtra::group_rows("US region", 2, 3, bold=FALSE) %>%
kableExtra::group_rows("Greeness", 4, 5, bold=FALSE) %>%
kableExtra::group_rows("US region", 7, 8, bold=FALSE) %>%
kableExtra::group_rows("Greeness", 9, 10, bold=FALSE) %>%
kableExtra::column_spec(2:6, width = "1cm") %>%
## kableExtra::kable_styling(latex_options = "striped") %>%
kableExtra::add_header_above(c(" "=1, "Monthly consumption\n(kBtu/sqft)" = 2, "# of days in a month \nwith mean temperature" = 2, "n" = 1), escape = TRUE) %>%
print()
sink()
}
allDataGreen %>%
dplyr::distinct(`Name`, `Organization`, `region_wiki`) %>%
dplyr::group_by(`Organization`, `region_wiki`) %>%
dplyr::summarise(n())
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
## end section: create summary of energy, weather, and building count for 4 alternative study set, based on different building types
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
## comparing static characteristics / balance of covariates of the two portfolio
load("../data/allDataGreen.rda")
balance.check.vars = c("GSF", "Sum of Employees", "Electric_(kBtu)",
"Gas_(kBtu)","eui_elec", "eui_gas", "eui_total", "HDD",
"CDD", "<10", "[20-30)", "[30-40)", "[40-50)" ,
"[50-60)", "[60-70)" , "[70-80)", "[80-90)" , ">90")
for (i in seq_along(folder.suffixes)) {
s = folder.suffixes[i]
print(s)
tokens = unlist(strsplit(s, "_"))
if (length(tokens) == 0) {
ownlease = NA
usetype = NA
} else {
usetype = paste0(tokens[2:3], collapse = "_")
if ("leased" %in% tokens) {
ownlease = NA
} else {
ownlease = "Owned"
}
}
if (is.na(ownlease)) {
own.or.lease.str = ""
} else {
own.or.lease.str = ownlease
}
result <- get.study.set(ownlease, usetype)
studyset = result$df
generic.office.gsa = result$generic.office.gsa
generic.office.pnc = result$generic.office.pnc
## may not be the properway to check whether the vector is NA
if (is.na(generic.office.gsa[1])) {
generic.office.gsa = "no restriction"
}
if (is.na(generic.office.pnc[1])) {
generic.office.pnc = "no restriction"
}
studyset %>%
distinct(Name, `type_general`, Organization) %>%
dplyr::group_by(Organization, `type_general`) %>%
dplyr::summarise(n()) %>%
print()
df.n = studyset %>%
dplyr::distinct(Organization, Name) %>%
dplyr::group_by(Organization) %>%
dplyr::summarise(cnt=n()) %>%
dplyr::ungroup() %>%
tibble::column_to_rownames("Organization") %>%
{.}
acc.balance = NULL
for (v in balance.check.vars) {
tresult <- t.test(as.formula(sprintf("`%s` ~ private", v)), data = studyset, alternative = "two.sided")
mu0 <- tresult$estimate[[1]]
mu1 <- tresult$estimate[[2]]
dif <- mu1 - mu0
p <- tresult$p.value
p.with.sig <- ifelse(p < 0.05, ifelse(p >= 0.01, sprintf("%.2f*", p), sprintf("%.2f**", p)), sprintf("%.2f", p))
print(p)
print(p.with.sig)
acc.balance <- rbind(acc.balance,
data.frame(variable = v, mu0 = mu0, mu1 = mu1, dif = dif, p = p, p.with.sig = p.with.sig))
}
sink(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/balance_compare%s.tex", s))
acc.balance %>%
dplyr::select(-p) %>%
knitr::kable("latex", booktabs = T,
format.args=list(big.mark=",", digits=2, scientific=F),
caption=sprintf("%s PNC (n=%s): %s; GSA (n=%s): %s",
own.or.lease.str,
df.n["PNC",1],
paste(generic.office.pnc, collapse = ", "),
df.n["GSA",1],
paste(generic.office.gsa, collapse = ", "))) %>%
kableExtra::kable_styling(full_width = FALSE, font_size = 7) %>%
kableExtra::group_rows("Static feature", 1, 2, bold=FALSE) %>%
kableExtra::group_rows("Energy", 3, 7, bold=FALSE) %>%
kableExtra::group_rows("Weather", 8, 18, bold=FALSE) %>%
## kableExtra::column_spec(2:6, width = "1cm") %>%
print()
sink()
}
dfstatic = allDataGreen %>%
dplyr::select(-`latitude`, -`longitude`, -`City`, -`State`, -`year`, -`Organization`, -`Date`) %>%
dplyr::mutate(Portfolio=ifelse(private==0, "public", "private")) %>%
{.}
## comparing building types of the two portfolio
num.building.per.type <- dfstatic %>%
dplyr::distinct(Name, Portfolio, `type_general`) %>%
dplyr::group_by(Portfolio, `type_general`) %>%
dplyr::summarise(n()) %>%
dplyr::ungroup() %>%
tidyr::spread(`Portfolio`, `n()`, fill=0L) %>%
dplyr::mutate(ordering = pmax(private, public)) %>%
dplyr::arrange(desc(ordering)) %>%
dplyr::select(-ordering) %>%
dplyr::mutate_at(vars(`type_general`), ~replace(., is.na(.), "Unknown")) %>%
{.}
result_table_tex <- xtable(num.building.per.type, caption="Building type count")
print(result_table_tex, tabular.environment = "tabular", include.rownames=FALSE,
file="~/Dropbox/thesis/writeups/policy_cmp/tables/num_building_per_type.tex")
## comparing static info of the two portfolio
s1 =
dfstatic %>%
dplyr::group_by(Portfolio) %>%
dplyr::summarise(GSF=mean(GSF), `Sum of Employees`=mean(`Sum of Employees`)) %>%
dplyr::ungroup() %>%
tibble::column_to_rownames("Portfolio") %>%
t() %>%
as.data.frame() %>%
tibble::rownames_to_column("variable") %>%
{.}
s2 =
dfstatic %>%
dplyr::mutate(Ownership = sprintf("%s percent", Ownership)) %>%
dplyr::group_by(Portfolio, Ownership) %>%
dplyr::summarise(count=n()) %>%
dplyr::ungroup() %>%
dplyr::group_by(Portfolio) %>%
dplyr::mutate(percentage=count/sum(count) * 100) %>%
dplyr::ungroup() %>%
dplyr::select(-count) %>%
tidyr::spread(Portfolio, `percentage`) %>%
dplyr::rename(variable=Ownership) %>%
{.}
s3 =
dfstatic %>%
dplyr::mutate(`Building Type`=ifelse(`Building Type` %in% c("Office", "Retail"), `Building Type`, "Other")) %>%
dplyr::group_by(Portfolio, `Building Type`) %>%
dplyr::summarise(count=n()) %>%
dplyr::ungroup() %>%
dplyr::group_by(Portfolio) %>%
dplyr::mutate(percentage = count/sum(count) * 100) %>%
dplyr::ungroup() %>%
dplyr::select(-count) %>%
tidyr::spread(Portfolio, `percentage`, fill=0) %>%
dplyr::rename(variable=`Building Type`) %>%
dplyr::arrange(desc(private)) %>%
{.}
t.test(dfstatic %>% dplyr::filter(Portfolio == "private") %>% .$GSF,
dfstatic %>% dplyr::filter(Portfolio == "public") %>% .$GSF)
static.cmp =
s1 %>%
bind_rows(s2) %>%
bind_rows(s3) %>%
dplyr::mutate(category=c("size", "", "Ownership", "", "", "Type", "", "")) %>%
dplyr::select(category, everything()) %>%
{.}
result_table_tex <- xtable(static.cmp, caption="Building static feature comparison")
print(result_table_tex, tabular.environment = "tabular", include.rownames=FALSE,
file="~/Dropbox/thesis/writeups/policy_cmp/tables/static_feature_cmp.tex")
allData %>%
dplyr::distinct(Organization, State, Name) %>%
dplyr::group_by(Organization, State) %>%
print()
## Plot GSA energy along with the policy change timeline
devtools::load_all("~/Dropbox/gsa_2017/db.interface")
gsa_energy =
db.interface::read_table_from_db(dbname = "all", tablename = "EUAS_monthly",
cols=c("Building_Number", "state_abbr", "Gross_Sq.Ft", "year", "month",
"Electric_(kBtu)", "Gas_(kBtu)", "Gross_Sq.Ft"
##, "Occupancy", "datacenter_sqft", "lab_sqft"
)) %>%
dplyr::rename(`Name`=`Building_Number`) %>%
## na.omit() %>%
{.}
## plot energy
toplot = gsa_energy %>%
dplyr::rename(`GSF`=`Gross_Sq.Ft`) %>%
dplyr::filter(`GSF`>0) %>%
dplyr::mutate(`Energy`=`Electric_(kBtu)` + `Gas_(kBtu)`) %>%
dplyr::group_by(year, month) %>%
dplyr::summarise(Energy=sum(Energy), GSF=sum(GSF)) %>%
dplyr::ungroup() %>%
## dplyr::mutate(target=Energy) %>%
dplyr::mutate(target=Energy/GSF) %>%
{.}
ylimit = max(toplot$target)
toplot %>%
dplyr::mutate(Time=zoo::as.yearmon(sprintf("%d-%d", year, month), format="%Y-%m")) %>%
ggplot2::ggplot(ggplot2::aes(x=Time, y=target)) +
ggplot2::geom_line() +
ggplot2::geom_vline(xintercept=as.yearmon("2005-08", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2005-08", format="%Y-%m"), y=1.0*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2005-08", format="%Y-%m"), y=1.0*ylimit, label="EPAct 2005") +
ggplot2::geom_vline(xintercept=as.yearmon("2006-01", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2006-01", format="%Y-%m"), y=0.9*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2006-01", format="%Y-%m"), y=0.9*ylimit, label="Guiding Principles\n(Memorandum in 2006)") +
ggplot2::geom_vline(xintercept=as.yearmon("2007-01", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2007-01", format="%Y-%m"), y=0.95*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2007-01", format="%Y-%m"), y=0.95*ylimit, label="Executive Order 13423") +
ggplot2::geom_vline(xintercept=as.yearmon("2007-12", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2007-12", format="%Y-%m"), y=1.0*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2007-12", format="%Y-%m"), y=1.0*ylimit, label="EISA 2007") +
ggplot2::geom_vline(xintercept=as.yearmon("2008-12", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2008-12", format="%Y-%m"), y=0.95*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2008-12", format="%Y-%m"), y=0.95*ylimit, label="Guiding Principles\n(revision in 2008)") +
ggplot2::geom_vline(xintercept=as.yearmon("2009-02", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2009-02", format="%Y-%m"), y=0.9*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2009-02", format="%Y-%m"), y=0.9*ylimit, label="Recovery Act") +
ggplot2::geom_vline(xintercept=as.yearmon("2009-10", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2009-10", format="%Y-%m"), y=1.0*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2009-10", format="%Y-%m"), y=1.0*ylimit, label="Executive Order 13514") +
ggplot2::geom_vline(xintercept=as.yearmon("2011-02", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2011-02", format="%Y-%m"), y=0.9*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2011-02", format="%Y-%m"), y=0.9*ylimit, label="President Memorandum") +
ggplot2::geom_vline(xintercept=as.yearmon("2011-06", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2011-06", format="%Y-%m"), y=0.95*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2011-06", format="%Y-%m"), y=0.95*ylimit, label="Better Buildings Initiative") +
ggplot2::geom_vline(xintercept=as.yearmon("2015-03", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2015-03", format="%Y-%m"), y=0.9*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2015-03", format="%Y-%m"), y=0.9*ylimit, label="Excecutive Order 13693") +
ggplot2::theme_bw() +
ggplot2::ylab("Electricity + Gas (kBtu/sqft)") +
ggplot2::ggtitle("The energy consumption trend and policy timelines of the public portfolio") +
ggplot2::theme()
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/gsa_eui_trend_policy_time.png", width=13, height=5, units="in")
## plot GSF
ylimit = max(toplot$GSF)
toplot %>%
dplyr::mutate(Time=zoo::as.yearmon(sprintf("%d-%d", year, month), format="%Y-%m")) %>%
ggplot2::ggplot(ggplot2::aes(x=Time, y=GSF)) +
ggplot2::geom_line() +
ggplot2::geom_vline(xintercept=as.yearmon("2010-06", format="%Y-%m"), linetype="dashed") +
ggplot2::annotate("text", x=as.yearmon("2010-06", format="%Y-%m"), y=0.9*ylimit, label="Memorandum June 2010\n(Disposing of Unneeded Federal Real Estate)") +
ggplot2::theme_bw() +
ggplot2::ggtitle("The GSF and time line the public portfolio") +
ggplot2::theme()
yujiex/pubPriCmp documentation built on Nov. 8, 2019, 8:05 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
library("dplyr")
library("readr")
library("pipeR")
library("feather")
library("ggplot2")
library("readxl")
library("xtable")
library("stringi")
library("zoo")
library("MASS")
library("kableExtra")
load("~/Dropbox/thesis/code/pubPriCmp/data/allData.rda")
load("~/Dropbox/thesis/code/pubPriCmp/data/allDataGreen.rda")
## EDA on buildings in East North Central region GSA portfolio
allDataGreen %>%
dplyr::select(`Name`, `Organization`, `eui_total`, `year`) %>%
dplyr::group_by(`Organization`, `Name`, `year`) %>%
dplyr::summarise(`eui_total`=sum(`eui_total`)) %>%
dplyr::ungroup() %>%
dplyr::group_by(`Organization`, `Name`) %>%
dplyr::summarise(`mean_annual_eui_total`=mean(`eui_total`)) %>%
dplyr::ungroup() %>%
ggplot2::ggplot() +
## ggplot2::geom_boxplot(ggplot2::aes(y=`mean_annual_eui_total`, x=`Organization`)) +
## ggplot2::ylim(c(0, 500)) +
ggplot2::geom_density(ggplot2::aes(x=`mean_annual_eui_total`, fill=`Organization`)) +
ggplot2::facet_wrap(.~`Organization`, ncol=1) +
ggplot2::theme()
allDataGreen %>%
dplyr::select(`Name`, `Organization`, `eui_total`, `year`, `month`) %>%
dplyr::group_by(`Organization`, `Name`, `year`) %>%
dplyr::summarise(`eui_total`=sum(`eui_total`)) %>%
dplyr::ungroup() %>%
dplyr::group_by(`Organization`, `Name`) %>%
dplyr::summarise(`mean_annual_eui_total`=mean(`eui_total`)) %>%
dplyr::ungroup() %>%
dplyr::filter(`mean_annual_eui_total` > 1000)
allDataGreen %>%
dplyr::select(`Name`, `Organization`, `eui_elec`, `eui_gas`, `year`, `month`, `GSF`) %>%
dplyr::filter(Name == "DC0001ZZ")
allDataGreen %>%
dplyr::distinct(`Name`, `GSF`) %>%
summary()
eastNorthCentral = allData %>%
dplyr::filter(`US Climate Region`=="East North Central") %>%
{.}
## plot energy trend
eastNorthCentral %>%
dplyr::mutate(`Date`=as.Date(sprintf("%s-%s-01", year, month), "%Y-%m-%d")) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
ggplot2::ggplot(ggplot2::aes(x=`Date`, y=`eui_elec_ln`, color=`Organization`, group=`Date`)) +
ggplot2::geom_violin() +
ggplot2::facet_wrap(.~`Organization`, ncol=1) +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.position = "bottom")
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/eastNorthCentral_distribution_by_time.png", width=6, height=4, units="in")
df_mean_avgminmax = feather::read_feather("building_mean_avgminmax.feather")
## plot public vs private portfolio average consumption by month
allData %>%
dplyr::filter(`Organization`=="GSA") %>%
dplyr::group_by(`Name`) %>%
dplyr::filter(sum(`>90`) > 0) %>%
dplyr::ungroup() %>%
dplyr::group_by(`Organization`, `Name`, `month`) %>%
dplyr::summarise(`eui_elec_ln`=mean(`eui_elec_ln`)) %>%
dplyr::ungroup() %>%
ggplot2::ggplot(ggplot2::aes(x=`month`, y=`eui_elec_ln`, color=`Organization`, group=`month`)) +
## ggplot2::geom_line() +
## ggplot2::geom_line() +
ggplot2::geom_boxplot() +
ggplot2::facet_wrap(.~`Organization`) +
ggplot2::theme_bw()
## plot individual building kernel smooth energy given mean monthly temperature
set.seed(0)
ratio = 0.2
number = 40
region = "East North Central"
## region = "Central"
## region = "Northeast"
## region = "Southeast"
samples =
allData %>%
dplyr::filter(`US Climate Region`==region) %>%
dplyr::select(`Name`, `Organization`) %>%
distinct(`Organization`, `Name`) %>%
dplyr::group_by(`Organization`) %>%
dplyr::slice(sample(1:n(), 30, replace=FALSE)) %>%
## dplyr::slice(sample(1:n(), as.integer(n()*ratio))) %>%
dplyr::ungroup() %>%
.$Name
## plot for all building
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::filter(`US Climate Region`==region) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
dplyr::group_by(`Name`) %>%
dplyr::mutate(`total_elec_eui`=sum(`eui_elec_ln`)) %>%
dplyr::ungroup() %>%
ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`, color=`total_elec_eui`)) +
## ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`)) +
ggplot2::geom_point(size=0.25) +
ggplot2::geom_smooth(span=1.0, se=FALSE, size=0.5) +
ggplot2::facet_wrap(~`Organization`) +
ggplot2::xlab("monthly average of daily mean temperature") +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::ggtitle(sprintf("kernel smoothing of building electricity (kBtu)/sqft on monthly mean temperature (%s)",
region)) +
ggplot2::theme_bw()
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_smooth_energy_avgtemp_all.png", gsub(" ", "", region)), width=6, height=4, units="in")
## plot a random sample of buildings
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::filter(`Name` %in% samples) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
dplyr::group_by(`Name`) %>%
dplyr::mutate(`total electricity per sqft`=sum(`eui_elec_ln`)) %>%
dplyr::ungroup() %>%
## ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`)) +
ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`, color=`total_elec_eui`)) +
ggplot2::geom_point(size=0.25) +
ggplot2::geom_smooth(span=1.0, se=FALSE, size=0.5) +
ggplot2::facet_wrap(~`Organization`) +
ggplot2::xlab("monthly average of daily mean temperature") +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::ggtitle(sprintf("kernel smoothing of building electricity (kBtu)/sqft on monthly mean temperature (%s)",
region)) +
ggplot2::theme_bw()
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_smooth_energy_avgtemp_sample.png", gsub(" ", "", region)), width=6, height=4, units="in")
## find buildings in the sample with a negative slope
downslope_buildings =
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::group_by(`Organization`, `Name`) %>%
dplyr::summarise(`slope`=lm(`eui_elec_ln`~`mean_avgminmax`)$coefficients[[2]]) %>%
dplyr::ungroup() %>%
dplyr::filter(`slope` < -0.005) %>%
.$Name
length(downslope_buildings)
## plot smoothed results for down slope buildings only
region = "East North Central"
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::filter(`Name` %in% downslope_buildings) %>%
dplyr::filter(`US Climate Region`==region) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`)) +
ggplot2::geom_point(size=0.25) +
ggplot2::geom_smooth(span=1.0, se=FALSE, size=0.5) +
ggplot2::facet_wrap(~`Organization`) +
ggplot2::xlab("monthly average of daily mean temperature") +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::ggtitle(sprintf("kernel smoothing of building electricity (kBtu)/sqft on monthly mean temperature (%s)\nwith negative slope", region)) +
ggplot2::theme_bw()
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_smooth_energy_avgtemp_downslope.png", gsub(" ", "", region)), width=6, height=4, units="in")
## find buildings in the sample with a negative slope
up_buildings =
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::group_by(`Organization`, `Name`) %>%
dplyr::summarise(`slope`=lm(`eui_elec_ln`~`mean_avgminmax`)$coefficients[[2]]) %>%
dplyr::ungroup() %>%
dplyr::filter(`slope` > 0.005) %>%
.$Name
length(up_buildings)
region = "East North Central"
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::filter(`Name` %in% up_buildings) %>%
dplyr::filter(`US Climate Region`==region) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
ggplot2::ggplot(ggplot2::aes(x=`mean_avgminmax`, y=`eui_elec_ln`, group=`Name`)) +
ggplot2::geom_point(size=0.25) +
ggplot2::geom_smooth(span=1.0, se=FALSE, size=0.5) +
ggplot2::facet_wrap(~`Organization`) +
ggplot2::xlab("monthly average of daily mean temperature") +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::ggtitle(sprintf("kernel smoothing of building electricity (kBtu)/sqft on monthly mean temperature (%s)\nwith negative slope", region)) +
ggplot2::theme_bw()
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_smooth_energy_avgtemp_up.png", gsub(" ", "", region)), width=6, height=4, units="in")
overall =
allData %>%
dplyr::left_join(df_mean_avgminmax, by=c("Name", "year", "month")) %>%
dplyr::select(`Name`, `Organization`, `eui_elec_ln`, `<10`, starts_with("["), `>90`) %>%
{.}
buildings = unique(overall$Name)
coef_labels = c("<10", sprintf("[%s-%s)", seq(10, 80, by=10), seq(20, 90, by=10)), ">90")
coef_labels <- coef_labels[which(coef_labels != "[60-70)")]
## transpose a data frame using the first column as the new header
transpose_df_header_1st_col <- function(df, colname) {
transposed = setNames(data.frame(t(df[,-1])), df[,1][, colname]) %>%
{.}
return(transposed)
}
acc = NULL
for (b in buildings) {
print(b)
df <- overall %>%
dplyr::filter(Name == b) %>%
## dplyr::mutate(`eui_elec_ln_scaled` = `eui_elec_ln` * 1e3) %>%
{.}
out <- lm(`eui_elec_ln`~`<10` + `[10-20)` + `[20-30)` + `[30-40)` +
## out <- lm(`eui_elec_ln_scaled`~`<10` + `[10-20)` + `[20-30)` + `[30-40)` +
`[40-50)` + `[50-60)` + `[70-80)` + `[80-90)` + `>90`, data=df)
coefs = as.numeric(out$coefficients)
coefs = coefs[2:length(coefs)]
names(coefs) = coef_labels
dfcoef = data.frame(coefs) %>%
tibble::rownames_to_column() %>%
dplyr::bind_rows(data.frame(rowname="[60-70)", coefs=NA)) %>%
dplyr::mutate(`Name`=b) %>%
{.}
acc <- rbind(acc, dfcoef)
}
acc %>%
readr::write_csv("bin_coef_by_building_scaleUp1e3.csv")
org_lookup = allData %>%
distinct(`Organization`, `US Climate Region`, `Name`)
total_eui_elec_lookup = allData %>%
dplyr::group_by(`Name`) %>%
dplyr::summarise(`total electricity per sqft`=sum(`eui_elec`)) %>%
dplyr::ungroup()
acc %>%
dplyr::filter(`rowname`=="<10", coefs < -2) %>%
## dplyr::filter(`rowname`=="<10", coefs < -2000) %>%
head()
## plot coefficient estimate for individual buildings for each region
acc %>%
dplyr::mutate(`ordering`=match(`rowname`, breaklabels)) %>%
dplyr::left_join(org_lookup) %>%
dplyr::left_join(total_eui_elec_lookup) %>%
dplyr::mutate(`US Climate Region`=factor(`US Climate Region`, levels=c("Central", "East North Central", "Northeast", "Southeast", "South"))) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
## dplyr::filter(`US Climate Region`=="East North Central") %>%
## dplyr::filter(`Name` != "MI1879ZZ") %>%
ggplot2::ggplot(ggplot2::aes(x=ordering, y=coefs, group=Name, color=`total electricity per sqft`)) +
ggplot2::geom_point(size=0.5) +
ggplot2::geom_line(size=0.5) +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::xlab("Temperature bin") +
ggplot2::scale_x_continuous(breaks=1:10, labels=breaklabels) +
ggplot2::facet_wrap(`Organization`~`US Climate Region`, nrow=2) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = ggplot2::element_text(angle=90, vjust=1),
legend.position = "bottom")
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/facet_building_bin_coef.png", width=8, height=6, units="in")
## restrict the plot range to -5000 to 5000
acc %>%
dplyr::mutate(`ordering`=match(`rowname`, breaklabels)) %>%
dplyr::left_join(org_lookup) %>%
dplyr::left_join(total_eui_elec_lookup) %>%
dplyr::mutate(`US Climate Region`=factor(`US Climate Region`, levels=c("Central", "East North Central", "Northeast", "Southeast", "South"))) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
dplyr::filter(coefs < 5, coefs > -5) %>%
## dplyr::filter(`total electricity per sqft`>100) %>%
ggplot2::ggplot(ggplot2::aes(x=ordering, y=coefs, group=Name, color=`total electricity per sqft`)) +
ggplot2::geom_point(size=0.5) +
ggplot2::geom_line(size=0.5) +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::xlab("Temperature bin") +
ggplot2::scale_x_continuous(breaks=1:10, labels=breaklabels) +
ggplot2::facet_wrap(`Organization`~`US Climate Region`, nrow=2) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = ggplot2::element_text(angle=90, vjust=1),
legend.position = "bottom")
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/facet_building_bin_coef_restrictYrange.png", width=8, height=6, units="in")
## count of days in 90 bin
num_days_in_bin_90 <- allData %>%
dplyr::select(`US Climate Region`, Name, `Organization`, `eui_elec_ln`, `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`temperature bin`, `count`, `<10`:`>90`) %>%
dplyr::filter(`temperature bin` %in% c(">90")) %>%
dplyr::filter(count > 0) %>%
dplyr::group_by(`US Climate Region`, `Organization`, `temperature bin`) %>%
dplyr::summarise(count = sum(count)) %>%
dplyr::ungroup() %>%
{.}
result_table_tex <- xtable(num_days_in_bin_90, caption="Count of days with daily temperature greater than 90")
print(result_table_tex, tabular.environment = "longtable", include.rownames=FALSE,
file="~/Dropbox/thesis/writeups/policy_cmp/tables/num_days_in_bin_90.tex")
head(num_days_in_bin_90)
## plot energy against weather for each region
## region = "East North Central"
## region = "Central"
## region = "Northeast"
## region = "Southeast"
region=""
to_plot = allData
if (nchar(region) > 0) {
to_plot = allData %>%
dplyr::filter(`US Climate Region`==region) %>%
{.}
}
to_plot %>%
dplyr::select(Name, `Organization`, `eui_elec_ln`, `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`temperature bin`, `count`, `<10`:`>90`) %>%
dplyr::filter(`temperature bin` %in% c("[80-90)", ">90")) %>%
ggplot2::ggplot(ggplot2::aes(x=`count`, y=`eui_elec_ln`, color=`Organization`)) +
ggplot2::geom_jitter(size=0.5, width=0.25) +
ggplot2::facet_wrap(`Organization`~`temperature bin`) +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::xlab("number of times per month daily mean temperature fall into this bin") +
ggplot2::ggtitle("Distribution of electricity use intensity by days in high temperature bins") +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_jitter_by_bin_cnt.png", gsub(" ", "", region)), width=6, height=4, units="in")
## plot energy distribution for bin 90
allData %>%
dplyr::select(`Name`, `US Climate Region`, `Organization`, `eui_elec_ln`, `>90`) %>%
dplyr::mutate(`US Climate Region`=factor(`US Climate Region`, levels=c("Central", "East North Central", "Northeast", "Southeast", "South"))) %>%
dplyr::mutate(`Organization`=recode(`Organization`, "GSA"="public", "PNC"="private")) %>%
dplyr::mutate(`Organization`=factor(`Organization`, levels=c("public", "private"))) %>%
ggplot2::ggplot(ggplot2::aes(x=`>90`, y=`eui_elec_ln`, color=`Organization`)) +
ggplot2::geom_jitter(size=0.5, width=0.25) +
ggplot2::ylab("Electricity in kBtu / gross sqft") +
ggplot2::xlab("number of times per month daily mean temperature > 90F") +
ggplot2::facet_wrap(`Organization`~`US Climate Region`, nrow=2) +
ggplot2::theme_bw() +
ggplot2::theme(legend.position = "bottom")
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/facet_energy_distribution_by_90bin.png",
width=6, height=4, units="in")
eastNorthCentral %>%
dplyr::group_by(`Organization`, `Name`) %>%
dplyr::slice(1) %>%
ungroup() %>%
dplyr::group_by(`Organization`) %>%
dplyr::summarise(n())
## plot histogram using full data
acc = feather::read_feather("~/Dropbox/thesis/code/pubPriCmp/data-raw/all_weather.feather")
## produce histogram of daily average temperature
lowerbound = min(acc$AVGMINMAX)
upperbound = max(acc$AVGMINMAX)
breaks = c(lowerbound, seq(10, 90, by=10), upperbound)
break_labels = c("<10", sprintf("[%s-%s)", seq(10, 80, by=10), seq(20, 90, by=10)), ">90")
## produce bin count plot
acc %>%
ggplot2::ggplot(ggplot2::aes(x=`AVGMINMAX`)) +
ggplot2::geom_histogram(breaks = breaks) +
ggplot2::stat_bin(breaks = breaks, geom="text", ggplot2::aes(label = ..count..), size=2, vjust=-1) +
ggtitle("Distribution of daily mean (average of min and max) temperature") +
xlab("daily mean temperature (average of min and max)") +
scale_x_continuous(breaks=seq(-30, 100, 10)) +
theme_bw()
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/avgminmax_bin.png", width=6, height=4, units="in")
## produce bar style count of daily average temp, as in Barreca et al. 2016
acc %>%
dplyr::mutate(`value_label`=dplyr::case_when(`AVGMINMAX` < 10 ~ "<10",
`AVGMINMAX` < 20 ~ break_labels[2],
`AVGMINMAX` < 30 ~ break_labels[3],
`AVGMINMAX` < 40 ~ break_labels[4],
`AVGMINMAX` < 50 ~ break_labels[5],
`AVGMINMAX` < 60 ~ break_labels[6],
`AVGMINMAX` < 70 ~ break_labels[7],
`AVGMINMAX` < 80 ~ break_labels[8],
`AVGMINMAX` < 90 ~ break_labels[9],
TRUE ~ break_labels[10]
)) %>%
dplyr::mutate(`value_label`=factor(`value_label`, levels=break_labels)) %>%
ggplot2::ggplot(ggplot2::aes(x=`value_label`)) +
ggplot2::geom_bar() +
ggtitle("Distribution of daily mean (average of min and max) temperature,\nof the three years") +
xlab("daily mean temperature (average of min and max)") +
ggplot2::theme_bw()
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/avgminmax_bin_barstyle.png", width=6, height=4, units="in")
## summary stats of weather data
devtools::load_all("~/Dropbox/thesis/code/summaryUtil")
summaryUtil::summary_table(acc) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_weather.csv")
acc_distance = feather::read_feather("~/Dropbox/thesis/code/pubPriCmp/data-raw/station_distance.feather")
## print number of distinct weather stations used in each year
acc_distance %>%
dplyr::group_by(`start_time`) %>%
dplyr::summarise(cnt = length(unique(`id`))) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/distinct_station_cnt_byyear.csv")
## print number of distinct weather stations used
acc_distance %>%
distinct(`id`) %>%
nrow()
distance_summary = acc_distance %>%
dplyr::group_by(`Name`, `varname`, `start_time`) %>%
dplyr::summarise(`minimum distance` = min(`distance`), `maximum distance`=max(`distance`), `average distance`=mean(`distance`), `number of stations`= n()) %>%
{.}
distance_summary %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_station_distance.csv")
distance_summary %>%
dplyr::filter(cnt < 3) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/fewerThan3Stations.csv")
## distribution of max distance of weather stations available
distance_summary %>%
dplyr::group_by(`Name`, `varname`) %>%
dplyr::summarise(`max_distance_overall`=max(`maximum distance`)) %>%
dplyr::ungroup() %>%
ggplot(aes(x=max_distance_overall)) +
geom_histogram() +
ggtitle("Distribution of maximum distance between weather station and building") +
xlab("maximum distance") +
facet_wrap(~ varname) +
theme(plot.title = element_text(size=12))
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/station_max_distance.png", width=6, height=4, units="in")
load("~/Dropbox/thesis/code/pubPriCmp/data/allData.rda")
load("~/Dropbox/thesis/code/pubPriCmp/data/allDataGreen.rda")
## generate appendix summary table for all variables
devtools::load_all("~/Dropbox/thesis/code/summaryUtil")
summaryUtil::summary_table(allData) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_allData.csv")
## generate appendix summary table for all variables with Greeness rating included
devtools::load_all("~/Dropbox/thesis/code/summaryUtil")
encodedGreen <- allDataGreen %>%
dplyr::mutate(`eGreeness`=ifelse(`Greeness` == "Least Green", 1, ifelse(`Greeness` == "Moderate Green", 2, 3))) %>%
{.}
unique(encodedGreen$eGreeness)
summaryUtil::summary_table(encodedGreen) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_allDataGreen.csv")
names(allDataGreen)
## generate in-paper summary statistics
yearlyDataStatic <-
allDataGreen %>%
dplyr::select(-`latitude`, -`longitude`, -`month`, -`City`, -`hasRetrofitBeforeThisMonth`, -`Date`, -`eui_elec_ln`,
-`eui_gas_ln`) %>%
dplyr::group_by(`Name`, `year`) %>%
dplyr::summarise_at(vars(`State`, `GSF`, `Organization`, `region_wiki`, `private`, `Greeness`), funs(first)) %>%
## dplyr::summarise_at(vars(`State`, `GSF`, `Organization`, `US Climate Region`, `private`, `Greeness`), funs(first)) %>%
{.}
names(yearlyDataStatic)
yearlyDataOther <-
allDataGreen %>%
dplyr::select(-`latitude`, -`longitude`, -`month`, -`City`, -`hasRetrofitBeforeThisMonth`, -`Date`, -`eui_elec_ln`,
-`eui_gas_ln`, -`US Climate Region`, -`USRegion`) %>%
## -`eui_gas_ln`) %>%
dplyr::group_by(`Name`, `year`) %>%
## dplyr::summarise_at(vars(-one_of("State", "GSF", "Organization", "US Climate Region", "private", "Greeness")),
dplyr::summarise_at(vars(-one_of("State", "GSF", "Organization", "region_wiki", "private", "Greeness")),
funs(sum)) %>%
{.}
names(yearlyDataOther)
yearlyData <- yearlyDataStatic %>%
dplyr::left_join(yearlyDataOther, by=c("Name", "year")) %>%
{.}
yearlyAvgStatic <- yearlyDataStatic %>%
dplyr::select(-`year`) %>%
dplyr::group_by(`Name`) %>%
dplyr::summarise_all(funs(first)) %>%
{.}
yearlyAvgOther <- yearlyDataOther %>%
dplyr::select(-`year`) %>%
dplyr::group_by(`Name`) %>%
dplyr::summarise_all(funs(mean)) %>%
{.}
yearlyAvgData <- yearlyAvgStatic %>%
dplyr::left_join(yearlyAvgOther, by="Name") %>%
{.}
head(yearlyAvgData)
## ## check whether additional rows are added as a result of the join
## nrow(yearlyAvgStatic)
## nrow(yearlyAvgData)
## bar plot of temperature bins, fill by ownership
break_labels = c("<10", sprintf("[%s-%s)", seq(10, 80, by=10), seq(20, 90, by=10)), ">90")
yearlyAvgData %>%
dplyr::select(-one_of("Name", "State", "Organization", "GSF", "HDD2", "CDD2", "US Climate Region", "USRegion",
"region_wiki", "eui_gas",
## dplyr::select(-one_of("Name", "State", "Organization", "GSF", "HDD2", "CDD2", "US Climate Region", "eui_gas",
"eui_elec", "HDD", "CDD", "Electric_(kBtu)", "Gas_(kBtu)")) %>%
tidyr::gather(`Mean Temperature Bin`, `Average Number of Days`, `<10`:`>90`) %>%
dplyr::mutate(`Mean Temperature Bin`=factor(`Mean Temperature Bin`, levels=break_labels)) %>%
dplyr::group_by(`private`, `Mean Temperature Bin`) %>%
dplyr::summarise(`Average Number of Days` = mean(`Average Number of Days`)) %>%
dplyr::mutate(`Ownership` = ifelse(`private`, "private", "public")) %>%
ggplot2::ggplot(ggplot2::aes(x=`Mean Temperature Bin`, y=`Average Number of Days`, fill=`Ownership`)) +
ggplot2::geom_bar(position = "dodge", stat="identity") +
ggplot2::scale_fill_grey() +
ggplot2::ggtitle("Distribution of average annual daily mean temperature") +
ggplot2::ylab("Number of days per year") +
ggplot2::theme_bw()
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/dailyAvgDistribution.png", width=6, height=4, units="in")
## bar plot of temperature bins, fill by Greeness, facet by public and private
break_labels = c("<10", sprintf("[%s-%s)", seq(10, 80, by=10), seq(20, 90, by=10)), ">90")
yearlyAvgData %>%
dplyr::select(-one_of("Name", "State", "Organization", "GSF", "HDD2", "CDD2", "US Climate Region", "USRegion",
"region_wiki", "eui_gas",
## dplyr::select(-one_of("Name", "State", "Organization", "GSF", "HDD2", "CDD2", "US Climate Region", "eui_gas",
"eui_elec", "HDD", "CDD", "Electric_(kBtu)", "Gas_(kBtu)")) %>%
tidyr::gather(`Mean Temperature Bin`, `Average Number of Days`, `<10`:`>90`) %>%
dplyr::mutate(`Mean Temperature Bin`=factor(`Mean Temperature Bin`, levels=break_labels),
`Greeness`=factor(`Greeness`, levels=c("Most Green", "Moderate Green", "Least Green"))) %>%
dplyr::group_by(`private`, `Greeness`, `Mean Temperature Bin`) %>%
dplyr::summarise(`Average Number of Days` = mean(`Average Number of Days`)) %>%
dplyr::mutate(`Ownership` = ifelse(`private`, "private", "public")) %>%
ggplot2::ggplot(ggplot2::aes(x=`Mean Temperature Bin`, y=`Average Number of Days`, fill=`Greeness`)) +
ggplot2::geom_bar(position = "dodge", stat="identity") +
ggplot2::scale_fill_grey() +
ggplot2::ggtitle("Distribution of daily average temperature") +
ggplot2::ylab("Number of days per year") +
ggplot2::facet_wrap(~`Ownership`, nrow=2) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1))
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/dailyAvgDistribution_greeness.png", width=6, height=6, units="in")
## bar plot of temperature bins, fill by region, facet by public and private
break_labels = c("<10", sprintf("[%s-%s)", seq(10, 80, by=10), seq(20, 90, by=10)), ">90")
yearlyAvgData %>%
dplyr::select(-one_of("Name", "State", "Organization", "GSF", "HDD2", "CDD2",
"US Climate Region", "USRegion", "Greeness",
"eui_gas",
"eui_elec", "HDD", "CDD", "Electric_(kBtu)",
"Gas_(kBtu)")) %>%
tidyr::gather(`Mean Temperature Bin`, `Average Number of Days`, `<10`:`>90`) %>%
dplyr::mutate(`Mean Temperature Bin`=factor(`Mean Temperature Bin`, levels=break_labels)) %>%
dplyr::group_by(`private`, `region_wiki`, `Mean Temperature Bin`) %>%
dplyr::summarise(`Average Number of Days` = mean(`Average Number of Days`)) %>%
dplyr::mutate(`Ownership` = ifelse(`private`, "private", "public")) %>%
ggplot2::ggplot(ggplot2::aes(x=`Mean Temperature Bin`, y=`Average Number of Days`, fill=`region_wiki`)) +
ggplot2::geom_bar(position = "dodge", stat="identity") +
ggplot2::scale_fill_grey() +
ggplot2::ggtitle("Distribution of daily average temperature") +
ggplot2::ylab("Number of days per year") +
ggplot2::facet_wrap(~`Ownership`, nrow=2) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1))
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/dailyAvgDistribution_region_wiki.png", width=6, height=6, units="in")
yearlyAvgData_cnt =
yearlyAvgData %>%
dplyr::group_by(`private`, `US Climate Region`) %>%
dplyr::summarise(cnt = n()) %>%
dplyr::ungroup() %>%
dplyr::mutate(`private`=as.character(`private`)) %>%
{.}
## summary by portfolio and climate region
yearlyAvgData %>%
dplyr::select(-`Greeness`) %>%
dplyr::mutate(`<40` = `<10` + `[10-20)` + `[20-30)` + `[30-40)`) %>%
dplyr::select(-one_of("Name", "State", "Organization", "<10", "[10-20)", "[20-30)", "[30-40)", "[40-50)",
"[50-60)", "[60-70)", "[70-80)")) %>%
dplyr::group_by(`private`, `US Climate Region`) %>%
dplyr::summarise_all(funs(mean)) %>%
dplyr::ungroup() %>%
dplyr::select(`private`, `US Climate Region`, `Electric_(kBtu)`, `Gas_(kBtu)`, `eui_elec`, `eui_gas`, `HDD`,
`CDD`, `<40`, `[80-90)`, `>90`) %>%
dplyr::mutate(`private`=as.character(`private`)) %>%
dplyr::mutate_if(is.numeric, function(x) {sprintf("%.1f", x)}) %>%
dplyr::left_join(yearlyAvgData_cnt, by=c("private", "US Climate Region")) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_annual_avg.csv")
## summary overall
yearlyAvgData %>%
dplyr::mutate(`<40` = `<10` + `[10-20)` + `[20-30)` + `[30-40)`) %>%
dplyr::select(-one_of("Name", "State", "Organization", "<10", "[10-20)", "[20-30)", "[30-40)", "[40-50)",
"[50-60)", "[60-70)", "[70-80)", "GSF", "US Climate Region", "private", "Greeness")) %>%
dplyr::summarise_all(mean) %>%
dplyr::select(`Electric_(kBtu)`, `Gas_(kBtu)`, `eui_elec`, `eui_gas`, `HDD`,
`CDD`, `<40`, `[80-90)`, `>90`) %>%
dplyr::mutate_if(is.numeric, function(x) {sprintf("%.1f", x)}) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_annual_avg_nogroup.csv")
## summary public and private
yearlyAvgData %>%
dplyr::mutate(`<40` = `<10` + `[10-20)` + `[20-30)` + `[30-40)`) %>%
dplyr::select(-one_of("Name", "State", "Organization", "<10", "[10-20)", "[20-30)", "[30-40)", "[40-50)",
"[50-60)", "[60-70)", "[70-80)", "GSF", "US Climate Region", "Greeness")) %>%
dplyr::group_by(`private`) %>%
dplyr::summarise_all(funs(mean)) %>%
dplyr::ungroup() %>%
dplyr::select(`private`, `Electric_(kBtu)`, `Gas_(kBtu)`, `eui_elec`, `eui_gas`, `HDD`,
`CDD`, `<40`, `[80-90)`, `>90`) %>%
dplyr::mutate(`private`=as.character(`private`)) %>%
dplyr::mutate_if(is.numeric, function(x) {sprintf("%.1f", x)}) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_annual_avg_pubpri.csv")
## summary by portfolio and greeness level
yearlyAvgData_cnt =
yearlyAvgData %>%
dplyr::group_by(`private`, `Greeness`) %>%
dplyr::summarise(cnt = n()) %>%
dplyr::ungroup()
yearlyAvgData %>%
dplyr::mutate(`<40` = `<10` + `[10-20)` + `[20-30)` + `[30-40)`) %>%
dplyr::select(-`US Climate Region`) %>%
dplyr::select(-one_of("Name", "State", "Organization", "<10", "[10-20)", "[20-30)", "[30-40)", "[40-50)",
"[50-60)", "[60-70)", "[70-80)")) %>%
dplyr::group_by(`private`, `Greeness`) %>%
dplyr::summarise_all(funs(mean)) %>%
dplyr::ungroup() %>%
dplyr::select(`private`, `Greeness`, `Electric_(kBtu)`, `Gas_(kBtu)`, `eui_elec`, `eui_gas`, `HDD`,
`CDD`, `<40`, `[80-90)`, `>90`) %>%
dplyr::mutate_if(is.numeric, function(x) {sprintf("%.1f", x)}) %>%
dplyr::mutate(`private`=as.numeric(`private`)) %>%
dplyr::left_join(yearlyAvgData_cnt, by=c("private", "Greeness")) %>%
readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_annual_avg_byGreeness.csv")
## produce plots of coefficients for temperature bin, with confidence interval
breaklevels = c("avgTempBelow10", "avgTemp10to20", "avgTemp20to30", "avgTemp30to40",
"avgTemp40to50", "avgTemp50to60", "avgTemp60to70", "avgTemp70to80", "avgTemp80to90",
"avgTempAbove90")
breaklabels = c("<10", "[10-20)", "[20-30)", "[30-40)",
"[40-50)", "[50-60)", "[60-70)", "[70-80)", "[80-90)",
">90")
load("../data/allData.rda")
## ------------------------------------------------------------------------------
## main analysis plots older version start
## ------------------------------------------------------------------------------
## portfolio = ""
## ## "_gas" is gas, "" is electricity
## energy_type = "_gas"
## status_code = ""
## df_zero = data.frame(X1="avgTemp60to70", variable=rep(c("coef", "ci_low", "ci_high"), 6),
## model=rep(c(rep("X2", 3), rep("X3", 3), rep("X4", 3)), 2),
## value = 0, status=c(rep("public", 9), rep("private relative to public", 9)))
## portfolio = "_GSA"
## ## energy_type = "_total"
## ## energy_type = "_gas"
## energy_type = ""
## status_code = "public"
## df_zero = data.frame(X1="avgTemp60to70", variable=rep(c("coef", "ci_low", "ci_high"), 2),
## model=rep(c(rep("X2", 3), rep("X3", 3)), 2),
## value = 0, status=c(rep("public", 6)))
portfolio = "_GSA"
## portfolio = "_PNC"
energy_type = "_total"
## energy_type = "_gas"
## energy_type = ""
## status_code = "private"
status_code = "public"
## office.class = "o_o"
## folder.suffix = paste0("_", office.class)
## office.class = ""
## folder.suffix = ""
for (portfolio in c("_GSA", "_PNC")) {
## could change this to other energy_type
for (ownership in c("owned", "owned_and_leased")) {
## energy_type = "_total"
for (energy_type in c("_gas")) {
## for (energy_type in c("", "_gas")) {
if (portfolio == "_PNC") {
status_code = "private"
} else if (portfolio == "_GSA") {
status_code = "public"
}
df_zero = data.frame(X1="avgTemp60to70", variable=rep(c("coef", "ci_low", "ci_high"), 2),
model=rep(c(rep("X2", 3), rep("X3", 3)), 2),
value = 0, status=c(rep(status_code, 6)))
for (office.class in c("", "o_o", "o_oct", "r_o", "ro_oct")) {
if (nchar(office.class) == 0) {
folder.suffix = ""
} else {
folder.suffix = paste0("_", office.class, "_", ownership)
}
## produce the bin coefficient and CI plot for the model with both portfolio
df =
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/reg_result%s/regression_results_bin_0720%s%s.txt", folder.suffix, tolower(portfolio), energy_type), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
tidyr::gather(`model`, `value`, starts_with("X")) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("0b.", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE),
!grepl("hasretrofit", `X1`, fixed = TRUE),
!grepl("Constant", `X1`, fixed = TRUE)
) %>%
{.}
if (status_code == "") {
df <- df %>%
dplyr::mutate(`status`=ifelse(grepl("1.private", X1, fixed=TRUE), "private relative to public", "public")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private relative to public"))) %>%
{.}
} else {
df <- df %>%
dplyr::mutate(`status`=status_code) %>%
{.}
}
df <- df %>%
dplyr::mutate(`X1`=gsub("1.private#c.", "", `X1`)) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
## readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/ci_bin.csv")
dplyr::bind_rows(df_zero) %>%
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="base line", "X3"="state specific trend", "X4"="state trend x private") %>%
{.}
if (nchar(portfolio) == 1) {
df <- df %>%
dplyr::filter(`model`=="state trend x private") %>%
{.}
} else {
df <- df %>%
dplyr::filter(`model`=="state specific trend") %>%
{.}
}
df <- df %>%
dplyr::filter(!(`X1` %in% c("avgTempBelow10", "avgTempAbove90"))) %>%
{.}
head(df)
min(df$value)
max(df$value)
if (energy_type == "") {
if (office.class == "") {
shifty = -8
yupperlimit = 15
}
if (ownership == "owned") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -30
yupperlimit = 40
} else if (office.class == "r_o") {
shifty = -8
yupperlimit = 15
} else if (office.class == "ro_oct") {
shifty = -8
yupperlimit = 15
}
} else if (ownership == "owned_and_leased") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -18
yupperlimit = 30
} else if (office.class == "r_o") {
shifty = -8
yupperlimit = 20
} else if (office.class == "ro_oct") {
shifty = -8
yupperlimit = 15
}
}
## shifty = -25
## yupperlimit = 25
scaling = 1 / 1e5 * 1.5
ylabel = "Electricity"
} else if (energy_type == "_gas") {
if (office.class == "") {
shifty = -10
yupperlimit = 55
}
if (ownership == "owned") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -40
yupperlimit = 75
} else if (office.class == "r_o") {
shifty = -10
yupperlimit = 70
} else if (office.class == "ro_oct") {
shifty = -10
yupperlimit = 60
}
} else if (ownership == "owned_and_leased") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -40
yupperlimit = 63
} else if (office.class == "r_o") {
shifty = -10
yupperlimit = 60
} else if (office.class == "ro_oct") {
shifty = -10
yupperlimit = 50
}
}
## shifty = -20
## yupperlimit = 40
scaling = 1 / 1e5 * 2
ylabel = "Gas"
} else if (energy_type == "_total") {
if (office.class == "") {
shifty = -10
yupperlimit = 40
}
if (ownership == "owned") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -30
yupperlimit = 60
} else if (office.class == "r_o") {
shifty = -10
yupperlimit = 45
} else if (office.class == "ro_oct") {
shifty = -10
yupperlimit = 40
}
} else if (ownership == "owned_and_leased") {
if (office.class %in% c("o_o", "o_oct")) {
shifty = -30
yupperlimit = 55
} else if (office.class == "r_o") {
shifty = -10
yupperlimit = 55
} else if (office.class == "ro_oct") {
shifty = -10
yupperlimit = 40
}
}
scaling = 1 / 1e5 * 2
ylabel = "Electricity + Gas"
}
temperatureBinData =
allData %>%
{.}
if (ownership == "owned") {
temperatureBinData <- temperatureBinData %>%
dplyr::filter(Ownership == "Owned")
}
if (office.class == "o_o") {
generic.office.pnc = "Office"
generic.office.gsa = "Office"
} else if (office.class == "o_oct") {
generic.office.pnc = "Office"
generic.office.gsa = c("Courthouse", "Office", "CT/Office")
} else if (office.class == "r_o") {
generic.office.pnc = "Retail"
generic.office.gsa = "Office"
} else if (office.class == "ro_oct") {
generic.office.pnc = c("Retail", "Office")
generic.office.gsa = c("Courthouse", "Office", "CT/Office")
}
if (office.class != "") {
temperatureBinData <- temperatureBinData %>%
dplyr::filter((Organization == "PNC" & `type_general` %in% generic.office.pnc) | (Organization == "GSA" & `type_general` %in% generic.office.gsa)) %>%
{.}
}
temperatureBinData <- temperatureBinData %>%
dplyr::select(`Organization`, `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
dplyr::group_by(`Organization`, `tempBin`) %>%
dplyr::summarise(count = sum(value)) %>%
dplyr::ungroup() %>%
dplyr::mutate(`count`=`count` * scaling) %>%
dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
{.}
if (status_code == "") {
temperatureBinData <- temperatureBinData %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private relative to public")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private relative to public"))) %>%
dplyr:::select(`count_start`, `count_end`, `ordering`, `status`) %>%
{.}
} else {
temperatureBinData <- temperatureBinData %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
dplyr:::select(`count_start`, `count_end`, `ordering`, `status`) %>%
{.}
}
head(temperatureBinData)
## dplyr::mutate_at(vars(variable), recode, "ci_low"="95% C.I. low end", "ci_high"="95% C.I. high end", "coef"="coefficient") %>%
df %>%
dplyr::left_join(temperatureBinData, by=c("status", "ordering")) %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value)) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:10, labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`, xend=`ordering`, y=`count_start`, yend=`count_end`), size=10) +
## ggplot2::geom_bar(mapping=ggplot2::aes(x=`ordering`, y=`count`), stat="identity", data=temperatureBinData) +
ggplot2::facet_wrap(`status`~`model`, ncol=1) +
ggplot2::xlab("Temperature Bin") +
ggplot2::ylab(sprintf("ln(%s in kBtu / gross sqft) * 1000", ylabel)) +
ggplot2::coord_cartesian(ylim=c(shifty, yupperlimit)) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci%s%s%s.png", portfolio, energy_type, folder.suffix), width=5, height=5, units="in")
}
}
}
}
## ------------------------------------------------------------------------------
## main analysis plots older version end
## ------------------------------------------------------------------------------
## ------------------------------------------------------------------------------
## main analysis plots newer version start
## ------------------------------------------------------------------------------
folder.suffixes = c("",
"_o_o_owned", "_o_oct_owned", "_r_o_owned", "_ro_oct_owned",
"_o_o_owned_and_leased", "_o_oct_owned_and_leased", "_r_o_owned_and_leased", "_ro_oct_owned_and_leased"
)
captions = c("no restrictions",
"Public: owned offices; Private: owned offices",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices",
"Public: owned offices; Private: owned retails",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices, retails",
"Public: offices; Private: offices",
"Public: offices, courthouses, CT/Offices; Private: offices",
"Public: offices; Private: retails",
"Public: offices, courthouses, CT/Offices; Private: offices, retails")
## energy_type = "_total"
energy_type = ""
## energy_type = "_gas"
for (modelsuffix in c("baseline", "statetrend")) {
for (i in seq_along(folder.suffixes)) {
s = folder.suffixes[i]
print(s)
tokens = unlist(strsplit(s, "_"))
if (length(tokens) == 0) {
ownlease = NA
usetype = NA
} else {
usetype = paste0(tokens[2:3], collapse = "_")
if ("leased" %in% tokens) {
ownlease = NA
} else {
ownlease = "Owned"
}
}
filename = sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/reg_result%s/regression_portfolio_%s%s.txt", s, modelsuffix,
energy_type)
fields_to_remove = c(".eDate", "eState", "Constant", "hasretrofitbeforethismonth", "0b.", "1b.eGreeness")
fields_to_keep = NULL
outreg_vars = c("coef", "se", "ci_low", "ci_high")
vars_to_keep = c("coef", "ci_low", "ci_high")
df <- read_outreg2(filename=filename, fields_to_remove= fields_to_remove, fields_to_keep= fields_to_keep,
outreg_vars=outreg_vars, vars_to_keep=vars_to_keep)
print(head(df))
model_info = tibble::tibble(`model_header`=sprintf("X%s", 2:(2 + 1)),
status=c("public", "private"))
df_zero = tibble::tibble(X1="avgTemp60to70",
variable=rep(c("coef", "ci_low", "ci_high"), 2),
model=modelsuffix,
value = 0,
status=c(rep("public", 3), rep("private", 3))) %>%
{.}
df <- df %>%
tidyr::gather(`model_header`, `value`, X2:(!!rlang::sym(sprintf("X%s", 2 + 1)))) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
dplyr::left_join(model_info, by="model_header") %>%
dplyr::mutate(`model`=modelname) %>%
dplyr::bind_rows(df_zero) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::select(-`model_header`) %>%
{.}
dfresult <- df %>%
dplyr::filter(!(`X1` %in% c("avgTempBelow10", "avgTempAbove90"))) %>%
{.}
lower = min(dfresult$value)
upper = max(dfresult$value)
temperatureBinData = get.study.set(ownlease, usetype)$df
temperatureSummary <- temperatureBinData %>%
dplyr::select(`Organization`, !!rlang::sym(regioncol), `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
dplyr::group_by(`Organization`, !!rlang::sym(regioncol), `tempBin`) %>%
dplyr::summarise(count = sum(value)) %>%
dplyr::ungroup() %>%
{.}
shifty = (-1)*(upper - lower)/2
scaling = abs(shifty) / max(temperatureSummary$count) / 2
temperatureToPlot <- temperatureSummary %>%
dplyr::mutate(`count`=`count` * scaling) %>%
dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
dplyr::select(`count_start`, `count_end`, `ordering`, `status`, !!rlang::sym(regioncol)) %>%
dplyr::rename(Rating = !!rlang::sym(regioncol)) %>%
{.}
xlabel = "Temperature Bin"
if (energy_type == "") {
ylabel = "Electricity"
} else if (energy_type == "_total") {
ylabel = "Electricity + Gas"
} else if (energy_type == "_gas") {
ylabel = "Gas"
}
## facet on public private and greeness level
dfresult %>%
dplyr::left_join(temperatureToPlot, by=c("ordering", "status")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private"))) %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value)) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`, xend=`ordering`, y=`count_start`, yend=`count_end`), size=7) +
## for slide
ggplot2::facet_wrap(.~`status`) +
## for paper
## ggplot2::facet_wrap(`Rating`~`status`, ncol=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(sprintf("ln(%s in kBtu / gross sqft) * 1000", ylabel)) +
## ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::labs(title=captions[i],
subtitle=(sprintf("%s model", modelsuffix))) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_separate_organization_%s_slide%s%s.png",
modelsuffix, energy_type, s), width=8, height=5, units="in")
}
}
## ------------------------------------------------------------------------------
## main analysis plots newer version start
## ------------------------------------------------------------------------------
breaklevels = c("hdd", "hdd2", "cdd", "cdd2")
breaklabels = c("HDD", "HDD2", "CDD", "CDD2")
xlabel = "Monthly HDD CDD"
## produce the hdd cdd model coefficient and CI plot for the model with both portfolio
readr::read_tsv("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_reg_0720.txt", skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`model`, `value`, X2:X7) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("0b.", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE),
!grepl("hasretrofit", `X1`, fixed = TRUE),
!grepl("Constant", `X1`, fixed = TRUE)
) %>%
dplyr::mutate(`status`=ifelse(grepl("1.private", X1, fixed=TRUE), "private relative to public", "public")) %>%
dplyr::mutate(`X1`=gsub("1.private#c.", "", `X1`)) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
## readr::write_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/ci_reg.csv")
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="baseline", "X3"="baseline (poly 2)", "X4" = "state specific trend",
"X5" = "state specific trend (poly 2)", "X6"="state trend x private",
"X7"="state trend x private (poly 2)") %>%
## dplyr::mutate_at(vars(variable), recode, "ci_low"="95% C.I. low end", "ci_high"="95% C.I. high end", "coef"="coefficient") %>%
dplyr::filter(`model` %in% c("state trend x private", "state trend x private (poly 2)")) %>%
tidyr::unite(`temp`, `X1`, `model`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "model"), sep="-") %>%
ggplot2::ggplot() +
ggplot2::geom_pointrange(ggplot2::aes(x=ordering, ymin=`ci_low`, ymax=`ci_high`, y=`coef`),
shape=18) +
## ggplot2::geom_errorbar(ggplot2::aes(x=ordering, ymin=`ci_low`, ymax=`ci_high`)) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::facet_wrap(`status`~`model`, nrow=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(NULL) +
ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank())
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/reg_coef_ci.png", width=5, height=5, units="in")
## portfolio = "GSA"
## status_code = "public"
portfolio = "PNC"
status_code = "private"
## produce the hdd cdd model coefficient and CI plot for the model for just one portfolio
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_reg_0720_%s.txt", portfolio), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`model`, `value`, X2:X5) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("0b.", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE),
!grepl("hasretrofit", `X1`, fixed = TRUE),
!grepl("Constant", `X1`, fixed = TRUE)
) %>%
dplyr::mutate(`status`=status_code) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
## readr::write_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/ci_reg_%s.csv", portfolio))
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="baseline", "X3"="baseline (poly 2)", "X4" = "state specific trend",
"X5" = "state specific trend (poly 2)") %>%
dplyr::filter(`model` %in% c("state specific trend", "state specific trend (poly 2)")) %>%
tidyr::unite(`temp`, `X1`, `model`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "model"), sep="-") %>%
ggplot2::ggplot() +
ggplot2::geom_pointrange(ggplot2::aes(x=ordering, ymin=`ci_low`, ymax=`ci_high`, y=`coef`), shape=18) +
## ggplot2::geom_errorbar(ggplot2::aes(x=ordering, ymin=`ci_low`, ymax=`ci_high`)) +
## ggplot2::geom_point(ggplot2::aes(x=ordering, y=coef)) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::facet_wrap(`status`~`model`, nrow=1) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(NULL) +
ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank())
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/reg_coef_ci_%s.png", portfolio), width=4, height=3, units="in")
breaklevels = c("avgTempBelow10", "avgTemp10to20", "avgTemp20to30", "avgTemp30to40",
"avgTemp40to50", "avgTemp50to60", "avgTemp60to70", "avgTemp70to80", "avgTemp80to90",
"avgTempAbove90")
breaklabels = c("<10", "[10-20)", "[20-30)", "[30-40)",
"[40-50)", "[50-60)", "[60-70)", "[70-80)", "[80-90)",
">90")
## produce coefficient plot of greeness for separate portfolio using error bars
portfolio = "GSA"
status_code = "public"
## portfolio = "PNC"
## status_code = "private"
xlabel = "Temperature Bin"
## produce the bin model coefficient and CI plot for greeness analysis for the model for just one portfolio
df <- readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_bin_greeness_0720_%s.txt", portfolio), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`model`, `value`, X2:X3) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("0b.", `X1`, fixed = TRUE),
!grepl("1b.eGreeness", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE),
!grepl("hasretrofit", `X1`, fixed = TRUE),
!grepl("Constant", `X1`, fixed = TRUE)
) %>%
dplyr::mutate(`status`=status_code) %>%
dplyr::mutate(`Rating`=ifelse(grepl("2.eGreeness", `X1`, fixed=TRUE), "Moderate Green \nrelative to Least Green",
ifelse(grepl("3.eGreeness", `X1`, fixed=TRUE), "Most Green \nrelative to Least Green", "Least Green"))) %>%
dplyr::mutate(`X1`=gsub("2.eGreeness#c.", "", `X1`)) %>%
dplyr::mutate(`X1`=gsub("3.eGreeness#c.", "", `X1`)) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="baseline", "X3" = "state specific trend") %>%
## readr::write_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/ci_bin_greeness_%s.csv", portfolio))
tidyr::unite(`temp`, `X1`, `model`, `Rating`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "model", "Rating"), sep="-") %>%
{.}
df %>%
ggplot2::ggplot() +
ggplot2::geom_errorbar(ggplot2::aes(x=ordering, ymin=`ci_low`, ymax=`ci_high`)) +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=coef)) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::facet_wrap(`Rating`~`model`, ncol=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(NULL) +
## ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank())
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness_%s.png", portfolio), width=8, height=5, units="in")
## produce coefficient plot of greeness for separate portfolio using connected dashed lines for conf int
df_zero = data.frame(X1="avgTemp60to70", variable=rep(c("coef", "ci_low", "ci_high"), 6),
model=rep(c(rep("X2", 9), rep("X3", 9)), 2),
Rating=rep(c(rep("Least Green", 3), rep("Moderate Green \nrelative to Least Green", 3),
rep("Most Green \nrelative to Least Green", 3)), 2),
value = 0, status=status_code)
## portfolio = "GSA"
## status_code = "public"
## energy_type = "_gas"
portfolio = "PNC"
status_code = "private"
energy_type = "_gas"
xlabel = "Temperature Bin"
## produce the bin model coefficient and CI plot for greeness analysis for the model for just one portfolio
df <- readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_bin_greeness_0720_%s%s.txt", portfolio, energy_type), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`model`, `value`, X2:X3) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("0b.", `X1`, fixed = TRUE),
!grepl("1b.eGreeness", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE),
!grepl("hasretrofit", `X1`, fixed = TRUE),
!grepl("Constant", `X1`, fixed = TRUE)
) %>%
dplyr::mutate(`status`=status_code) %>%
dplyr::mutate(`Rating`=ifelse(grepl("2.eGreeness", `X1`, fixed=TRUE), "Moderate Green \nrelative to Least Green",
ifelse(grepl("3.eGreeness", `X1`, fixed=TRUE), "Most Green \nrelative to Least Green", "Least Green"))) %>%
dplyr::mutate(`X1`=gsub("2.eGreeness#c.", "", `X1`)) %>%
dplyr::mutate(`X1`=gsub("3.eGreeness#c.", "", `X1`)) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
dplyr::bind_rows(df_zero) %>%
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="baseline", "X3" = "state specific trend") %>%
## readr::write_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/ci_bin_greeness_%s.csv", portfolio))
{.}
min(df$value)
max(df$value)
if (energy_type == "") {
shifty = -45
yupperlimit = 25
scaling = 1 / 1e4 * 1.5
} else if (energy_type == "_gas") {
if (status_code == "public") {
shifty = -260
} else {
shifty = -100
}
yupperlimit = 75
scaling = 1 / 1e4 * 5
}
temperatureBinData =
allDataGreen %>%
{.}
temperatureBinData <-
temperatureBinData %>%
dplyr::select(`Organization`, `Greeness`, `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
dplyr::group_by(`Organization`, `Greeness`, `tempBin`) %>%
dplyr::summarise(count = sum(value)) %>%
dplyr::ungroup() %>%
dplyr::mutate(`count`=`count` * scaling) %>%
dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
{.}
if (status_code == "") {
temperatureBinData <- temperatureBinData %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private relative to public")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private relative to public"))) %>%
dplyr:::select(`count_start`, `count_end`, `ordering`, `status`, `Greeness`) %>%
{.}
} else {
temperatureBinData <- temperatureBinData %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
dplyr:::select(`count_start`, `count_end`, `ordering`, `status`, `Greeness`) %>%
{.}
}
temperatureBinData <- temperatureBinData %>%
dplyr::mutate(`Rating`=recode(Greeness, "Moderate Green"="Moderate Green \nrelative to Least Green",
"Most Green"="Most Green \nrelative to Least Green")) %>%
dplyr::select(-`Greeness`) %>%
{.}
df %>%
dplyr::left_join(temperatureBinData, by=c("ordering", "status", "Rating")) %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value)) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`, xend=`ordering`, y=`count_start`, yend=`count_end`), size=7) +
## for slide
ggplot2::facet_wrap(`model`~`Rating`, ncol=3) +
## for paper
## ggplot2::facet_wrap(`Rating`~`model`, ncol=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab("ln(Electricity in kBtu / gross sqft) * 1000") +
## ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness_%s_slide%s.png", portfolio, energy_type), width=8, height=5, units="in")
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness_%s.png", portfolio), width=6, height=7, units="in")
## produce coefficient plot of greeness for separate portfolio by greeness level
## using connected dashed lines for conf int
read_outreg2 <- function(filename, fields_to_remove, fields_to_keep, outreg_vars,
vars_to_keep) {
df =
readr::read_tsv(filename, skip=7, col_names=FALSE) %>%
head(-5) %>%
dplyr::mutate(`variable`=rep(outreg_vars, nrow(.)/length(outreg_vars))) %>%
tidyr::fill(`X1`) %>%
dplyr::filter(`variable` %in% vars_to_keep) %>%
{.}
for (field in fields_to_remove) {
df <- df %>%
dplyr::filter(!grepl(field, `X1`, fixed = TRUE)) %>%
{.}
}
for (field in fields_to_keep) {
df <- df %>%
dplyr::filter(grepl(field, `X1`, fixed = TRUE)) %>%
{.}
}
return(df)
}
breaklevels = c("avgTempBelow10", "avgTemp10to20", "avgTemp20to30", "avgTemp30to40",
"avgTemp40to50", "avgTemp50to60", "avgTemp60to70", "avgTemp70to80", "avgTemp80to90",
"avgTempAbove90")
breaklabels = c("<10", "[10-20)", "[20-30)", "[30-40)",
"[40-50)", "[50-60)", "[60-70)", "[70-80)", "[80-90)",
">90")
## green_levels = c("Least Green", "Moderate Green", "Most Green")
green_levels = c("Least Green", "Most Green")
len_green_levels = length(green_levels)
## energy_type = "_gas"
## modelsuffix = "baseline"
## modelname = "baseline"
## energy_type = "_gas"
energy_type = "_total"
modelsuffix = "statetrend"
modelname = "state specific trend"
## regioncol = "Greeness"
## filenamesuffix = ""
## green.source = "Majority Vote"
## regioncol = "greeness.howe.state"
## filenamesuffix = "_howestate"
## green.source = "State Data from Howe 2015"
regioncol = "greeness.howe.county"
filenamesuffix = "_howecounty"
green.source = "County Data from Howe 2015"
folder.suffixes = c("",
"_o_o_owned", "_o_oct_owned", "_r_o_owned", "_ro_oct_owned",
"_o_o_owned_and_leased", "_o_oct_owned_and_leased", "_r_o_owned_and_leased", "_ro_oct_owned_and_leased"
)
captions = c("no restrictions",
"Public: owned offices; Private: owned offices",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices",
"Public: owned offices; Private: owned retails",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices, retails",
"Public: offices; Private: offices",
"Public: offices, courthouses, CT/Offices; Private: offices",
"Public: offices; Private: retails",
"Public: offices, courthouses, CT/Offices; Private: offices, retails")
for (i in seq_along(folder.suffixes)) {
s = folder.suffixes[i]
print(s)
tokens = unlist(strsplit(s, "_"))
if (length(tokens) == 0) {
ownlease = NA
usetype = NA
} else {
usetype = paste0(tokens[2:3], collapse = "_")
if ("leased" %in% tokens) {
ownlease = NA
} else {
ownlease = "Owned"
}
}
filename = sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/reg_result%s/regression_greeness_%s%s%s.txt", s, modelsuffix,
energy_type, filenamesuffix)
fields_to_remove = c(".eDate", "eState", "Constant", "hasretrofitbeforethismonth", "0b.", "1b.eGreeness")
fields_to_keep = NULL
outreg_vars = c("coef", "se", "ci_low", "ci_high")
vars_to_keep = c("coef", "ci_low", "ci_high")
df <- read_outreg2(filename=filename, fields_to_remove= fields_to_remove, fields_to_keep= fields_to_keep,
outreg_vars=outreg_vars, vars_to_keep=vars_to_keep)
head(df)
model_info = tibble::tibble(`model_header`=sprintf("X%s", 2:(2*len_green_levels + 1)),
status=c(rep("public", len_green_levels), rep("private", len_green_levels)),
Rating=rep(green_levels, 2))
df_zero = tibble::tibble(X1="avgTemp60to70",
variable=rep(c("coef", "ci_low", "ci_high"), 2 * len_green_levels),
model=modelname,
Rating=rep(sort(rep(green_levels, 3)), 2), value = 0,
status=c(rep("public", 3 * len_green_levels), rep("private", 3 * len_green_levels))) %>%
{.}
df <- df %>%
tidyr::gather(`model_header`, `value`, X2:(!!rlang::sym(sprintf("X%s", 2 * len_green_levels + 1)))) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
dplyr::left_join(model_info, by="model_header") %>%
dplyr::mutate(`model`=modelname) %>%
dplyr::bind_rows(df_zero) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::select(-`model_header`) %>%
{.}
dfresult <- df %>%
dplyr::filter(!(`X1` %in% c("avgTempBelow10", "avgTempAbove90"))) %>%
{.}
lower = min(dfresult$value)
upper = max(dfresult$value)
temperatureBinData = get.study.set(ownlease, usetype)$df
temperatureSummary <- temperatureBinData %>%
dplyr::select(`Organization`, !!rlang::sym(regioncol), `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
dplyr::group_by(`Organization`, !!rlang::sym(regioncol), `tempBin`) %>%
dplyr::summarise(count = sum(value)) %>%
dplyr::ungroup() %>%
{.}
shifty = (-1)*(upper - lower)/2
scaling = abs(shifty) / max(temperatureSummary$count) / 2
temperatureToPlot <- temperatureSummary %>%
dplyr::mutate(`count`=`count` * scaling) %>%
dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
dplyr::select(`count_start`, `count_end`, `ordering`, `status`, !!rlang::sym(regioncol)) %>%
dplyr::rename(Rating = !!rlang::sym(regioncol)) %>%
{.}
## if (energy_type == "") {
## shifty = -45
## yupperlimit = 25
## scaling = 1 / 1e4 * 1.5
## ylabel = "Electricity"
## } else if (energy_type == "_gas") {
## shifty = -45
## yupperlimit = 240
## scaling = 1 / 1e4 * 1.5
## ylabel = "Gas"
## } else if (energy_type == "_total") {
## shifty = -5
## yupperlimit = 40
## scaling = 1 / 1e5 * 1.5
## ylabel = "Electricity + Gas"
## }
## temperatureBinData =
## allDataGreen %>%
## {.}
## temperatureBinData <-
## temperatureBinData %>%
## dplyr::select(`Organization`, `Greeness`, `<10`, starts_with("["), `>90`) %>%
## tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
## dplyr::group_by(`Organization`, `Greeness`, `tempBin`) %>%
## dplyr::summarise(count = sum(value)) %>%
## dplyr::ungroup() %>%
## dplyr::mutate(`count`=`count` * scaling) %>%
## dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
## dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
## dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
## dplyr::select(`count_start`, `count_end`, `ordering`, `status`, `Greeness`) %>%
## dplyr::rename(`Rating`=`Greeness`) %>%
## {.}
## head(temperatureBinData)
## head(df)
xlabel = "Temperature Bin"
ylabel = "Electricity + Gas"
## facet on public private and greeness level
dfresult %>%
dplyr::left_join(temperatureToPlot, by=c("ordering", "status", "Rating")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private"))) %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value)) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`, xend=`ordering`, y=`count_start`, yend=`count_end`), size=7) +
## for slide
ggplot2::facet_wrap(`status`~`Rating`, ncol=len_green_levels) +
## for paper
## ggplot2::facet_wrap(`Rating`~`status`, ncol=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(sprintf("ln(%s in kBtu / gross sqft) * 1000", ylabel)) +
## ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::labs(title=captions[i],
subtitle=(sprintf("%s model (greeness source: %s)", modelsuffix, green.source))) +
ggplot2::theme_bw() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness%s_separate_organization_%s_slide%s%s.png",
filenamesuffix, modelsuffix, energy_type, s), width=8, height=5, units="in")
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness_separate_organization_greeness_%s.png", modelsuffix), width=6, height=7, units="in")
## facet on public private and fill on greeness level
to_plot <- dfresult %>%
dplyr::left_join(temperatureToPlot, by=c("ordering", "status", "Rating")) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private"))) %>%
{.}
to_plot_least <- to_plot %>%
dplyr::filter(`Rating`=="Least Green") %>%
{.}
to_plot_most <- to_plot %>%
dplyr::filter(`Rating`=="Most Green") %>%
{.}
binsize = 4
jitterAmount = 0.15
xlabel = "Temperature Bin"
## facet on public private and greeness level
to_plot %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value, color=Rating)) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable, color=Rating)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
## ggplot2::geom_segment(ggplot2::aes(x=`ordering`, xend=`ordering`, y=`count_start`, yend=`count_end`), size=7) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`-jitterAmount, xend=`ordering`-jitterAmount, y=`count_start`, yend=`count_end`, group=`Rating`, color=`Rating`), data=to_plot_least, size=binsize) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`+jitterAmount, xend=`ordering`+jitterAmount, y=`count_start`, yend=`count_end`, group=`status`, color=`Rating`), data=to_plot_most, size=binsize) +
## for slide
ggplot2::facet_wrap(.~`status`, ncol=2) +
## for paper
## ggplot2::facet_wrap(`Rating`~`status`, ncol=2) +
ggplot2::xlab(xlabel) +
ggplot2::ylab(sprintf("ln(%s in kBtu / gross sqft) * 1000", ylabel)) +
## ggplot2::coord_cartesian(ylim=c(-300, 750)) +
ggplot2::theme_bw() +
ggplot2::labs(title=captions[i],
subtitle=(sprintf("%s model (greeness source: %s)", modelsuffix, green.source))) +
ggplot2::scale_color_manual(values=c("#A1D99B", "#31A354")) +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/facet_pubpri_bin_coef_ci_greeness%s_separate_organization_%s_slide%s%s.png", filenamesuffix, modelsuffix, energy_type, s), width=8, height=5, units="in")
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_greeness_separate_organization_greeness_%s.png", modelsuffix), width=6, height=7, units="in")
}
allDataGreen %>%
distinct(Organization, Greeness, Name) %>%
dplyr::group_by(Organization, Greeness) %>%
dplyr::summarise(n())
## produce coefficient plot by climate region for separate portfolio using connected dashed lines for conf int
breaklevels = c("avgTempBelow10", "avgTemp10to20", "avgTemp20to30", "avgTemp30to40",
"avgTemp40to50", "avgTemp50to60", "avgTemp60to70", "avgTemp70to80", "avgTemp80to90",
"avgTempAbove90")
breaklabels = c("<10", "[10-20)", "[20-30)", "[30-40)",
"[40-50)", "[50-60)", "[60-70)", "[70-80)", "[80-90)",
">90")
modelname = "baseline"
xlabel = "Temperature Bin"
process_climate_result <- function(modelname, modellabel, energy_type, region_keyword, folder_suffix) {
## produce the bin model coefficient and CI plot for greeness analysis for the model for just one portfolio
df_recode_xi =
readr::read_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/header_to_%s_owner.csv", region_keyword)) %>%
{.}
filename = sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/reg_result%s/regression_%s_%s%s.txt",
folder_suffix, region_keyword, modelname, energy_type)
print(filename)
fields_to_remove = c("0b.", "eDate", "eState", "hasretrofit", "Constant", "o.avgTempBelow10")
fields_to_keep = NULL
outreg_vars = c("coef", "se", "ci_low", "ci_high")
vars_to_keep = c("coef", "ci_low", "ci_high")
df_model = read_outreg2(filename=filename, fields_to_remove=fields_to_remove, fields_to_keep=fields_to_keep,
outreg_vars=outreg_vars, vars_to_keep=vars_to_keep)
## df_model =
## readr::read_tsv(),
## skip=7, col_names=FALSE) %>%
## ## remove bottom non-conformative rows
## head(-5) %>%
## dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
## tidyr::fill(`X1`) %>%
## tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## ## modify this to X2:X?, ? is the number of models + 1
## tidyr::gather(`header`, `value`, X2:X9) %>%
## tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
## dplyr::filter(`variable` != "se",
## !grepl("0b.", `X1`, fixed = TRUE),
## !grepl("eDate", `X1`, fixed = TRUE),
## !grepl("eState", `X1`, fixed = TRUE),
## !grepl("hasretrofit", `X1`, fixed = TRUE),
## !grepl("Constant", `X1`, fixed = TRUE),
## !grepl("o.avgTempBelow10", `X1`, fixed = TRUE)
## ) %>%
df_zero = data.frame(X1="avgTemp60to70", variable=rep(c("coef", "ci_low", "ci_high"), 8),
header=sprintf("X%s", sort(rep(2:(length(unique(df_recode_xi$header)) + 1), 3))), value=0)
print(df_zero, n=Inf)
df_model <- df_model %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
tidyr::gather(`header`, `value`, starts_with("X")) %>%
## tidyr::gather(`header`, `value`, X2:X9) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=as.numeric(`value`)) %>%
na.omit() %>%
dplyr::bind_rows(df_zero) %>%
dplyr::mutate(`X1`=factor(X1, levels=breaklevels)) %>%
dplyr::mutate(`ordering`=match(`X1`, breaklevels)) %>%
dplyr::left_join(df_recode_xi, by="header") %>%
dplyr::mutate(`model`=modellabel) %>%
{.}
return(df_model)
}
breaklevels = c("avgTempBelow10", "avgTemp10to20", "avgTemp20to30", "avgTemp30to40",
"avgTemp40to50", "avgTemp50to60", "avgTemp60to70", "avgTemp70to80", "avgTemp80to90",
"avgTempAbove90")
breaklabels = c("<10", "[10-20)", "[20-30)", "[30-40)",
"[40-50)", "[50-60)", "[60-70)", "[70-80)", "[80-90)",
">90")
## energy_type = "_gas" ## use "" for electricity
energy_type = "_total"
## region_keyword = "climateregion"
## region_keyword = "usregion"
region_keyword = "region_wiki"
## regioncol = "US Climate Region"
## regioncol = "USRegion"
regioncol = "region_wiki"
## modelname = "statetrend"
## modelname = "baseline"
modelname = "both"
folder.suffixes = c("",
"_o_o_owned", "_o_oct_owned", "_r_o_owned", "_ro_oct_owned",
"_o_o_owned_and_leased", "_o_oct_owned_and_leased", "_r_o_owned_and_leased", "_ro_oct_owned_and_leased"
)
captions = c("no restrictions",
"Public: owned offices; Private: owned offices",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices",
"Public: owned offices; Private: owned retails",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices, retails",
"Public: offices; Private: offices",
"Public: offices, courthouses, CT/Offices; Private: offices",
"Public: offices; Private: retails",
"Public: offices, courthouses, CT/Offices; Private: offices, retails")
for (i in seq_along(folder.suffixes)) {
i = 1
s = folder.suffixes[i]
print(s)
tokens = unlist(strsplit(s, "_"))
if (length(tokens) == 0) {
ownlease = NA
usetype = NA
} else {
usetype = paste0(tokens[2:3], collapse = "_")
if ("leased" %in% tokens) {
ownlease = NA
} else {
ownlease = "Owned"
}
}
dfresult =
process_climate_result(modelname="baseline", modellabel="baseline", energy_type=energy_type,
region_keyword=region_keyword, s) %>%
dplyr::bind_rows(process_climate_result(modelname="statetrend", modellabel="state specific trend",
energy_type=energy_type,
region_keyword=region_keyword, s)) %>%
{.}
## remove 10F and 90F bins
dfresult <- dfresult %>%
dplyr::filter(!(`X1` %in% c("avgTempBelow10", "avgTempAbove90"))) %>%
{.}
## head(dfresult)
lower = min(dfresult$value)
upper = max(dfresult$value)
## (dfresult) %>%
## distinct(!!rlang::sym(regioncol), `status`, `model`)
## if (energy_type == "") {
## shifty=-400
## yupperlimit = 750
## scaling = 1 / 1e3 * 1.5
## } else if (energy_type == "_gas") {
## shifty=-900
## yupperlimit = 900
## scaling = 1 / 1e3 * 1.5
## } else if (energy_type == "_total") {
## ## shifty=-130
## ## yupperlimit = 220
## ## scaling = 1 / 1e4 * 4
## shifty=-50
## yupperlimit = 80
## scaling = 1 / 1e4 * 1.5
## }
temperatureBinData = get.study.set(ownlease, usetype)$df
temperatureSummary <- temperatureBinData %>%
dplyr::select(`Organization`, !!rlang::sym(regioncol), `<10`, starts_with("["), `>90`) %>%
tidyr::gather(`tempBin`, `value`, `<10`:`>90`) %>%
dplyr::group_by(`Organization`, !!rlang::sym(regioncol), `tempBin`) %>%
dplyr::summarise(count = sum(value)) %>%
dplyr::ungroup() %>%
{.}
shifty = (-1)*(upper - lower)/2
scaling = abs(shifty) / max(temperatureSummary$count) / 2
temperatureToPlot <- temperatureSummary %>%
dplyr::mutate(`count`=`count` * scaling) %>%
dplyr::mutate(`count_start`=shifty, `count_end`=`count` + shifty) %>%
dplyr::mutate(`ordering`=match(`tempBin`, breaklabels)) %>%
dplyr::mutate(`status`=ifelse(`Organization`=="GSA", "public", "private")) %>%
dplyr::select(`count_start`, `count_end`, `ordering`, `status`, !!rlang::sym(regioncol)) %>%
{.}
head(temperatureToPlot)
to_plot <- dfresult %>%
dplyr::left_join(temperatureToPlot, by=c("ordering", "status", regioncol)) %>%
dplyr::mutate(`status`=factor(`status`, levels=c("public", "private"))) %>%
{.}
to_plot_public <- to_plot %>%
dplyr::filter(`status`=="public") %>%
{.}
to_plot_private <- to_plot %>%
dplyr::filter(`status`=="private") %>%
{.}
to_plot_north <- to_plot %>%
dplyr::filter(!!rlang::sym(regioncol)=="North") %>%
{.}
to_plot_south <- to_plot %>%
dplyr::filter(!!rlang::sym(regioncol)=="South") %>%
{.}
## for paper
## binsize = 4
## jitterAmount = 0.3
## for slides
binsize = 1.7
jitterAmount = 0.15
if (energy_type == "_total") {
binsize = 3.4
jitterAmount = 0.17
}
## facet by portfolio
p <- to_plot %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value, color=!!rlang::sym(regioncol)), size=0.7) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable, color=!!rlang::sym(regioncol))) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`-jitterAmount, xend=`ordering`-jitterAmount, y=`count_start`, yend=`count_end`, group=!!rlang::sym(regioncol), color=!!rlang::sym(regioncol)), data=to_plot_north, size=binsize) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`+jitterAmount, xend=`ordering`+jitterAmount, y=`count_start`, yend=`count_end`, group=!!rlang::sym(regioncol), color=!!rlang::sym(regioncol)), data=to_plot_south, size=binsize) +
## for paper
## ggplot2::facet_wrap(`US Climate Region`~`model`, ncol=2) +
## for slides
## use nrow=2 for climate region plots
ggplot2::facet_wrap(as.formula(paste("model~", "status")), ncol=2) +
## ggplot2::facet_wrap(`model`~`USRegion`, nrow=2) +
ggplot2::xlab("Temperature Bin") +
ggplot2::ylab("ln(Electricity in kBtu / gross sqft) * 1000") +
## ggplot2::coord_cartesian(ylim=c(shifty, yupperlimit)) +
## this is to plot slide_zoom for electric
## ggplot2::ylim(c(-100, 100)) +
## this is to plot slide_zoom for gas
## ggplot2::ylim(c(-150, 150)) +
ggplot2::theme_bw() +
ggplot2::scale_color_manual(values = RColorBrewer::brewer.pal(3, "RdBu")[c(3, 1)], name="US Region") +
ggplot2::ggtitle(captions[i]) +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
if (regioncol == "US Climate Region") {
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate_slide_zoom%s.png", energy_type), width=8, height=5, units="in")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate_slide%s.png", energy_type), width=8, height=5, units="in")
## for paper
## ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate.png", width=8, height=7, units="in")
}
else {
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_%s_slide_zoom%s.png", regioncol, energy_type), width=8, height=5, units="in")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_%s_slide%s_%s%s.png", regioncol, energy_type, modelname, s), width=8, height=6, units="in")
## for paper
## ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_usregion.png", width=8, height=7, units="in")
}
## facet by region
p <- to_plot %>%
ggplot2::ggplot() +
ggplot2::geom_point(ggplot2::aes(x=ordering, y=value, color=status), size=0.7) +
ggplot2::geom_line(ggplot2::aes(x=ordering, y=value, linetype=variable, color=status)) +
scale_linetype_manual(values=c("dashed", "dashed", "solid"), breaks=c("ci_low", NA, "coef"), labels=c("95% C.I.", "", "Coefficient")) +
scale_x_continuous(breaks=1:length(breaklabels), labels=breaklabels) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`-jitterAmount, xend=`ordering`-jitterAmount, y=`count_start`, yend=`count_end`, group=!!rlang::sym("status"), color=!!rlang::sym("status")), data=to_plot_public, size=binsize) +
ggplot2::geom_segment(ggplot2::aes(x=`ordering`+jitterAmount, xend=`ordering`+jitterAmount, y=`count_start`, yend=`count_end`, group=!!rlang::sym("status"), color=!!rlang::sym("status")), data=to_plot_private, size=binsize) +
## for paper
## ggplot2::facet_wrap(`US Climate Region`~`model`, ncol=2) +
## for slides
## use nrow=2 for climate region plots
ggplot2::facet_wrap(as.formula(paste("model~", regioncol)), ncol=2) +
## ggplot2::facet_wrap(`model`~`USRegion`, nrow=2) +
ggplot2::xlab("Temperature Bin") +
ggplot2::ylab("ln(Electricity in kBtu / gross sqft) * 1000") +
## ggplot2::coord_cartesian(ylim=c(shifty, yupperlimit)) +
## this is to plot slide_zoom for electric
## ggplot2::ylim(c(-100, 100)) +
## this is to plot slide_zoom for gas
## ggplot2::ylim(c(-150, 150)) +
ggplot2::ggtitle(captions[i]) +
ggplot2::theme_bw() +
ggplot2::scale_color_grey() +
ggplot2::theme(axis.text.x = element_text(angle = 90, hjust = 1),
legend.title=element_blank(), legend.position = "bottom")
print(p)
if (regioncol == "US Climate Region") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate_slide_zoom%s_facetRegion.png", energy_type), width=8, height=5, units="in")
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate_slide%s_facetRegion.png", energy_type), width=8, height=5, units="in")
## for paper
## ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_climate_facetRegion.png", width=8, height=7, units="in")
}
else {
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_%s_slide_zoom%s_facetRegion.png", regioncol, energy_type), width=8, height=5, units="in")
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_%s_slide%s_%s_facetRegion%s.png", regioncol, energy_type, modelname, s), width=8, height=5, units="in")
## for paper
## ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/bin_coef_ci_usregion_facetRegion.png", width=8, height=7, units="in")
}
}
## generate image tex file
newlines = NULL
## imagePrefix = "bin_coef_ci_region_wiki_slide_total_both_facetRegion"
## filename = "facetRegionPlots"
## imagePrefix = "bin_coef_ci_region_wiki_slide_total_both"
## filename = "facetPortfolioPlots"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_separate_organization_baseline_slide_total"
## filename = "facetGreenessPlots_baseline_majority"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_howestate_separate_organization_baseline_slide_total"
## filename = "facetGreenessPlots_baseline_howestate"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_howecounty_separate_organization_baseline_slide_total"
## filename = "facetGreenessPlots_baseline_howecounty"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_separate_organization_statetrend_slide_total"
## filename = "facetGreenessPlots_statetrend_majority"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_howestate_separate_organization_statetrend_slide_total"
## filename = "facetGreenessPlots_statetrend_howestate"
## imagePrefix = "facet_pubpri_bin_coef_ci_greeness_howecounty_separate_organization_statetrend_slide_total"
## filename = "facetGreenessPlots_statetrend_howecounty"
## imagePrefix = "bin_coef_ci_separate_organization_baseline_slide_total"
## filename = "main_analysis_baseline_total"
imagePrefix = "bin_coef_ci_separate_organization_baseline_slide_gas"
filename = "main_analysis_baseline_gas"
## imagePrefix = "bin_coef_ci_separate_organization_baseline_slide"
## filename = "main_analysis_baseline"
for (s in folder.suffixes) {
newlines <- c(newlines,
"\\begin{frame}",
"\\begin{figure}[h!]",
"\\centering",
sprintf("\\includegraphics[width=1.0\\linewidth]{images/%s%s.png}",
imagePrefix, s),
"\\end{figure}",
"\\end{frame}")
}
con <- file(sprintf("~/Dropbox/thesis/writeups/policy_cmp/images/%s.tex", filename), open = "w+")
writeLines(newlines, con, sep = "\n", useBytes = FALSE)
close(con)
## generate tex order by building type first then by greenness source
for (modeltype in c("baseline", "statetrend")) {
newlines = NULL
filename = sprintf("facetGreenessPlots_%s", modeltype)
for (s in folder.suffixes) {
for (green.suffix in c("", "_howestate", "_howecounty")) {
imagePrefix = sprintf("facet_pubpri_bin_coef_ci_greeness%s_separate_organization_%s_slide_total",
green.suffix, modeltype)
newlines <- c(newlines,
"\\begin{frame}",
"\\begin{figure}[h!]",
"\\centering",
sprintf("\\includegraphics[width=1.0\\linewidth]{images/%s%s.png}",
imagePrefix, s),
"\\end{figure}",
"\\end{frame}")
}
con <- file(sprintf("~/Dropbox/thesis/writeups/policy_cmp/images/%s.tex", filename), open = "w+")
writeLines(newlines, con, sep = "\n", useBytes = FALSE)
close(con)
}
}
## generate regression result summary table for bin model with greeness
## portfolio="PNC"
## status_code = "private"
portfolio="GSA"
status_code = "public"
summary_portfolio = readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_bin_greeness_0720_%s.txt", portfolio), skip=7, col_names=FALSE) %>%
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
dplyr::filter(!(`variable` %in% c("ci_low", "ci_high")),
!grepl("1b.eGreeness", `X1`, fixed = TRUE),
!grepl("eDate", `X1`, fixed = TRUE),
!grepl("eState", `X1`, fixed = TRUE)
) %>%
dplyr::mutate(`X1`=gsub("2.eGreeness#c.", "Moderate Green x ", `X1`)) %>%
dplyr::mutate(`X1`=gsub("3.eGreeness#c.", "Most Green x ", `X1`)) %>%
dplyr::mutate(`X1`=ifelse(as.numeric(rownames(.)) %% 2 == 1, `X1`, "")) %>%
{.}
tail_summary_portfolio = readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/regression_results_bin_greeness_0720_%s.txt", portfolio), skip=7, col_names=FALSE) %>%
tail(5) %>%
{.}
summary_export = summary_portfolio %>%
dplyr::bind_rows(tail_summary_portfolio) %>%
{.}
summary_export %>%
readr::write_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_greeness_model_%s.csv", portfolio))
summary_export <-
summary_export %>%
dplyr::select(-`variable`) %>%
dplyr::rename(" "=`X1`, "(1)"=`X2`, "(2)"=`X3`) %>%
{.}
result_table_tex <- xtable(summary_export, caption=sprintf("Regression result for the %s portfolio", status_code))
print(result_table_tex, tabular.environment = "longtable", include.rownames=FALSE,
file=sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_greeness_result.tex", portfolio))
# long table of state temperature
avg_temp_state =
readr::read_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/stateAverageTemperature2011to2013.csv") %>%
tibble::as_data_frame() %>%
dplyr::select(`Location`, `Value`) %>%
dplyr::rename(`Average temperature (F)`=`Value`, `State`=`Location`) %>%
{.}
result_table_tex <- xtable(avg_temp_state, caption="Average state temperature from 2011 to 2013")
print(result_table_tex, tabular.environment = "longtable", include.rownames=FALSE,
file="~/Dropbox/thesis/writeups/policy_cmp/tables/avg_state_temp.tex", size="\\footnotesize")
## only plot the temperature bin model
models = "Temperature Bin"
plotHeight = 6
## plot all models
## models=c("No Control", "HDD CDD", "Polynomial degree 2 HDD and CDD", "Temperature Bin")
## warmth_class_method = "building temperature"
## warmth_class_method = "state latitude"
warmth_class_method = "state temperature"
## plot coefficients for retrofit climate event study
pd <- position_dodge(width = 0.4)
## quarterly plot setting starts -------------------------
duration = "quarterly"
energy_type = ""
## energy_type = "_gas"
endbound = 20
fatten_amount = 2.5
size = 0.2
point_size = 1.0
## quarterly plot setting ends -------------------------
## monthly plot setting starts -------------------------
## duration = "monthly"
## energy_type = "_gas"
## endbound = 60
## fatten_amount = 2.5
## size = 0.2
## point_size = 1.0
## monthly plot setting ends -------------------------
## yearly plot setting starts -------------------------
## duration = "yearly"
## endbound = 5
## fatten_amount = 2.5
## size = 1.0
## point_size = 2.5
## yearly plot setting ends -------------------------
retrofits = c("Advanced Metering", "Building Envelope", "HVAC", "Lighting", "GSALink", "Bundle")
names(retrofits) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
if (energy_type == "") {
energy_label = "Electricity"
} else if (energy_type == "_gas") {
energy_label = "Gas"
}
## for (fileSuffix in c("metering")) {
for (fileSuffix in c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")) {
retrofit = retrofits[[fileSuffix]]
if (warmth_class_method == "building temperature") {
p <-
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_buildingTemp_event_study_%s%s.txt", duration, fileSuffix, energy_type), skip=7, col_names=FALSE)
} else if (warmth_class_method == "state temperature") {
p <-
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_event_study_%s%s.txt", duration, fileSuffix, energy_type), skip=7, col_names=FALSE)
} else if (warmth_class_method == "state latitude") {
p <-
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_stateLat_event_study_%s%s.txt", duration, fileSuffix, energy_type), skip=7, col_names=FALSE)
}
## remove bottom non-conformative rows
p <- p %>%
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`model`, `value`, X2:X5) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl("o.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("0b.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("3b.eWarmth", `X1`, fixed = TRUE),
!grepl("o.eWarmth", `X1`, fixed = TRUE),
!grepl("cdd", `X1`, fixed=TRUE),
!grepl("hdd", `X1`, fixed=TRUE),
!grepl("Constant", `X1`, fixed=TRUE),
!grepl("avgTemp", `X1`, fixed=TRUE),
) %>%
{.}
if (duration == "year") {
p <- p %>%
dplyr::filter(!grepl(".year", `X1`, fixed = TRUE)) %>%
{.}
} else if ((duration == "monthly") | (duration == "quarterly")){
print("asdf")
p <- p %>%
dplyr::filter(!grepl(".eDate", `X1`, fixed = TRUE)) %>%
{.}
}
p <- p %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=gsub(",", "", `value`)) %>%
tidyr::unite(`temp`, `X1`, `model`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "model"), sep="-") %>%
dplyr::filter(!is.na(`coef`)) %>%
## readr::write_csv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_coef_ci.csv", fileSuffix))
dplyr::mutate(`coef`=as.numeric(`coef`),
`ci_high`=as.numeric(`ci_high`),
`ci_low`=as.numeric(`ci_low`),
) %>%
dplyr::mutate(warmth = ifelse(grepl("1.eWarmth", X1, fixed=TRUE), "cold",
ifelse(grepl("2.eWarmth", X1, fixed=TRUE), "hot", "mild"))) %>%
dplyr::mutate(`tau`=as.integer(substr(X1, 1, stringi::stri_locate(X1, fixed=".", mode="first")-1))) %>%
dplyr::mutate(`tau`=ifelse(`tau` <= endbound, `tau`, `tau` - 2*endbound - 1)) %>%
dplyr::mutate(`warmth`=ifelse(as.numeric(tau) < 0, "pre-retrofit all warmth level", `warmth`)) %>%
dplyr::mutate_at(vars(model), recode, "X2"="No Control", "X3"="HDD CDD", "X4"="Polynomial degree 2 HDD and CDD",
"X5"="Temperature Bin") %>%
dplyr::select(-`X1`) %>%
dplyr::bind_rows(data.frame(ci_low=NA, ci_high=NA, coef=0, term=NA,
warmth=rep(c("cold", "hot", "mild", "pre-retrofit all warmth level"), 4),
model=c(rep("No Control", 4), rep("HDD CDD", 4),
rep("Polynomial degree 2 HDD and CDD", 4), rep("Temperature Bin", 4)),
tau=0)) %>%
{.}
print(head(p))
p <- p %>%
dplyr::filter(`model` %in% models) %>%
{.}
p <- p %>%
ggplot2::ggplot(ggplot2::aes(x=tau, y=coef, color=warmth, group=warmth)) +
ggplot2::geom_line() +
## ggplot2::geom_point() +
ggplot2::geom_pointrange(ggplot2::aes(x=tau, y=`coef`, ymin=`ci_low`, ymax=`ci_high`, color=warmth, group=warmth), position=pd, shape=18, fatten=fatten_amount, size=size) +
ggplot2::geom_point(x=0, y=0, shape=18, size=point_size) +
ggplot2::facet_wrap(`model`~., ncol=2) +
ggplot2::xlab("Number of years before of after retrofit") +
ggplot2::ylab(NULL) +
## ggplot2::coord_cartesian(ylim=c(-20, 25)) +
## ggplot2::scale_color_brewer(palette="Spectral") +
## ggplot2::scale_color_brewer(palette="RdBu") +
## ggplot2::scale_color_manual(values=c("#5698C4", "#E97550", "grey")) +
## ggplot2::scale_color_manual(values=c("#d55e00", "#8B8CB0", "skyblue")) +
## ggplot2::scale_color_manual(values=c("#d55e00", "#009e73", "skyblue")) +
## logan's
## ggplot2::scale_color_manual(values=colorRampPalette(c("#FF0000","#000066"))(3L)) +
ggplot2::scale_color_manual(values=c("#67A9CF", "#EF8A62", "grey70", "black")) +
## ggplot2::theme_dark() +
ggplot2::theme_bw() +
ggtitle(sprintf("%s Event study \"%s\" retrofit\nWarmth level by %s",
energy_label, retrofit, warmth_class_method)) +
ggplot2::theme(legend.position = "bottom",
plot.title = element_text(size=slide_size))
print(p)
## ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_retrofit_coef_noci_%s.png", duration, fileSuffix), width=8, height=5, units="in")
if (warmth_class_method == "building temperature") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_buildingTemp_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type), width=8, height=plotHeight, units="in")
} else if (warmth_class_method == "state temperature") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type), width=8, height=plotHeight, units="in")
} else if (warmth_class_method == "state latitude") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_stateLat_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type), width=8, height=plotHeight, units="in")
}
}
df_recode_retrofit =
readr::read_csv("~/Dropbox/thesis/code/pubPriCmp/data-raw/separate_model_monthly_header_to_retrofit.csv") %>%
{.}
## plot coefficients for retrofit climate event study with each warmth level fitted separately
## plot all models
## models = c("no_control", "hddcdd", "hdd2cdd2", "bin")
## plotHeight = 5
## plot one model
models = "bin"
plotHeight = 6
slide_size = 22
## warmth_class_method = "state temperature"
## warmth_class_method = "state latitude"
warmth_class_method = "building temperature"
pd <- position_dodge(width = 0.4)
## quarterly plot setting starts -------------------------
duration = "quarterly"
## energy_type = ""
energy_type = "_gas"
endbound = 20
fatten_amount = 2.5
size = 0.2
point_size = 1.0
## quarterly plot setting ends -------------------------
## monthly plot setting starts -------------------------
## duration = "monthly"
## energy_type = "_gas"
## endbound = 60
## fatten_amount = 2.5
## size = 0.2
## point_size = 1.0
## monthly plot setting ends -------------------------
## yearly plot setting starts -------------------------
## duration = "yearly"
## endbound = 5
## fatten_amount = 2.5
## size = 1.0
## point_size = 2.5
## yearly plot setting ends -------------------------
if (energy_type == "") {
energy_label = "Electricity"
} else if (energy_type == "_gas") {
energy_label = "Gas"
}
acc <- lapply(models, function(model_type) {
if (warmth_class_method == "state temperature") {
dfacc =
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/separate_model_%s_retrofit_climate_%s%s.txt", duration, model_type, energy_type), skip=7, col_names=FALSE)
} else if (warmth_class_method == "state latitude") {
dfacc =
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/separate_model_%s_stateLat_retrofit_climate_%s%s.txt", duration, model_type, energy_type), skip=7, col_names=FALSE)
} else if (warmth_class_method == "building temperature") {
dfacc =
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/separate_model_%s_buildingTemp_retrofit_climate_%s%s.txt", duration, model_type, energy_type), skip=7, col_names=FALSE)
}
dfacc <- dfacc %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`header`, `value`, X2:X19) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl(".eDate", `X1`, fixed = TRUE),
!grepl("o.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("0b.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("cdd", `X1`, fixed=TRUE),
!grepl("hdd", `X1`, fixed=TRUE),
!grepl("Constant", `X1`, fixed=TRUE),
!grepl("avgTemp", `X1`, fixed=TRUE),
) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=gsub(",", "", `value`)) %>%
tidyr::unite(`temp`, `X1`, `header`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "header"), sep="-") %>%
dplyr::filter(!is.na(`coef`)) %>%
dplyr::mutate(`coef`=as.numeric(`coef`),
`ci_high`=as.numeric(`ci_high`),
`ci_low`=as.numeric(`ci_low`),
) %>%
dplyr::mutate(`tau`=as.integer(substr(X1, 1, stringi::stri_locate(X1, fixed=".", mode="first")-1))) %>%
dplyr::mutate(`tau`=ifelse(`tau` <= endbound, `tau`, `tau` - 2*endbound - 1)) %>%
dplyr::left_join(df_recode_retrofit, by="header") %>%
dplyr::mutate(`model`=model_type) %>%
dplyr::select(-`X1`, -`header`) %>%
{.}
})
modelResult = do.call(rbind, acc)
retrofits = c("Advanced Metering", "Building Envelope", "HVAC", "Lighting", "GSALink", "Bundle")
names(retrofits) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
for (fileSuffix in names(retrofits)) {
retrofit_type <- retrofits[[fileSuffix]]
print(retrofit_type)
p <- modelResult %>%
dplyr::bind_rows(data.frame(`tau`=0, `ci_low`=0, `ci_high`=0, `coef`=0, `retrofit`=retrofit_type,
`warmth`=rep(c("mild", "cold", "hot"), 4),
`model`=c(rep("no_control", 3), rep("hddcdd", 3), rep("hdd2cdd2", 3),
rep("bin", 3)))) %>%
dplyr::filter(`model` %in% models) %>%
dplyr::filter(retrofit == retrofit_type) %>%
ggplot2::ggplot(ggplot2::aes(x=tau, y=coef, color=warmth, group=warmth)) +
ggplot2::geom_line() +
ggplot2::geom_pointrange(ggplot2::aes(x=tau, y=`coef`, ymin=`ci_low`, ymax=`ci_high`, color=warmth, group=warmth), position=pd, shape=18, fatten=fatten_amount, size=size) +
ggplot2::geom_point(x=0, y=0, shape=18, size=point_size) +
ggplot2::facet_wrap(`model`~., ncol=2) +
ggplot2::xlab("Number of years before of after retrofit") +
ggplot2::ylab(NULL) +
ggplot2::scale_color_manual(values=c("#67A9CF", "#EF8A62", "grey70", "black")) +
ggplot2::theme_bw() +
ggtitle(sprintf("%s Event study \"%s\" retrofit\nWarmth level by %s", energy_label, retrofit_type, warmth_class_method)) +
ggplot2::theme(legend.position = "bottom",
plot.title = element_text(size=slide_size))
print(p)
if (warmth_class_method == "state temperature") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/sep_%s_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type),width=8, height=plotHeight, units="in")
} else if (warmth_class_method == "state latitude") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/sep_%s_stateLat_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type), width=8, height=plotHeight, units="in")
} else if (warmth_class_method == "building temperature") {
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/sep_%s_buildingTemp_retrofit_coef_ci_%s_%s%s.png", duration, fileSuffix, length(models), energy_type), width=8, height=plotHeight, units="in")
}
}
## generate summary statistics tables
retrofits = c("Advanced Metering", "Building Envelope", "HVAC", "Lighting", "GSALink", "")
names(retrofits) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
## for (fileSuffix in c("metering")) {
for (fileSuffix in c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")) {
retrofit = retrofits[[fileSuffix]]
summary_head = readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/yearly_event_study_%s.txt", fileSuffix), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
dplyr::filter(!(`variable` %in% c("ci_low", "ci_high"))) %>%
dplyr::filter(
## `variable` != "se",
!grepl(".year", `X1`, fixed = TRUE),
!grepl("o.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("0b.tau_bounded_post", `X1`, fixed = TRUE),
!grepl("3b.eWarmth", `X1`, fixed = TRUE),
!grepl("o.eWarmth", `X1`, fixed = TRUE),
) %>%
dplyr::mutate(X1 = gsub("1.eWarmth", "cold", X1)) %>%
dplyr::mutate(X1 = gsub("2.eWarmth", "hot", X1)) %>%
dplyr::mutate(X1 = gsub("1.tau_bounded_post", "$I[\\\\tau=1]$", X1)) %>%
dplyr::mutate(X1 = gsub("2.tau_bounded_post", "$I[\\\\tau=2]$", X1)) %>%
dplyr::mutate(X1 = gsub("3.tau_bounded_post", "$I[\\\\tau=3]$", X1)) %>%
dplyr::mutate(X1 = gsub("4.tau_bounded_post", "$I[\\\\tau=4]$", X1)) %>%
dplyr::mutate(X1 = gsub("5.tau_bounded_post", "$I[\\\\tau>=5]$", X1)) %>%
dplyr::mutate(X1 = gsub("1.tau_bounded_shift", "$I[\\\\tau=1]$", X1)) %>%
dplyr::mutate(X1 = gsub("2.tau_bounded_shift", "$I[\\\\tau=2]$", X1)) %>%
dplyr::mutate(X1 = gsub("3.tau_bounded_shift", "$I[\\\\tau=3]$", X1)) %>%
dplyr::mutate(X1 = gsub("4.tau_bounded_shift", "$I[\\\\tau=4]$", X1)) %>%
dplyr::mutate(X1 = gsub("5.tau_bounded_shift", "$I[\\\\tau=5]$", X1)) %>%
dplyr::mutate(X1 = gsub("6.tau_bounded_shift", "$I[\\\\tau<=-5]$", X1)) %>%
dplyr::mutate(X1 = gsub("7.tau_bounded_shift", "$I[\\\\tau=-4]$", X1)) %>%
dplyr::mutate(X1 = gsub("8.tau_bounded_shift", "$I[\\\\tau=-3]$", X1)) %>%
dplyr::mutate(X1 = gsub("9.tau_bounded_shift", "$I[\\\\tau=-2]$", X1)) %>%
dplyr::mutate(X1 = gsub("10.tau_bounded_shift", "$[\\\\tau=-1]$", X1)) %>%
dplyr::mutate(X1 = gsub("#", "$\\\\times$", X1)) %>%
dplyr::mutate(`X1`=ifelse(as.numeric(rownames(.)) %% 2 == 1, `X1`, "")) %>%
## dplyr::select(-`variable`) %>%
{.}
print(head(summary_head))
summary_tail = readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/yearly_event_study_%s.txt", fileSuffix), skip=7, col_names=FALSE) %>%
tail(5) %>%
dplyr::mutate(`X1` = gsub(" ci_high", "", `X1`)) %>%
dplyr::mutate(`X1` = gsub("<", "$<$", `X1`)) %>%
{.}
print(head(summary_tail))
summary_export = summary_head %>%
dplyr::bind_rows(summary_tail) %>%
{.}
summary_export %>%
readr::write_csv(sprintf("~/Dropbox/thesis/code/pubPriCmp/data-raw/summary_retrofit_model_%s.csv", fileSuffix))
summary_export <-
summary_export %>%
dplyr::select(-`variable`) %>%
dplyr::rename(" "=`X1`, "(1)"=`X2`, "(2)"=`X3`, "(3)"=`X4`, "(4)"=`X5`) %>%
{.}
result_table_tex <- xtable(summary_export, caption=sprintf("Regression result for event study model of buildings with %s retrofit", fileSuffix))
print(result_table_tex, tabular.environment = "longtable", include.rownames=FALSE,
file=sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/summary_retrofit_model_%s.tex", fileSuffix), size="\\footnotesize", sanitize.text.function=function(x){x})
## above sanitize function overwritten so that the math symbols stay the original way
}
## plot coefficients for retrofit climate event study with each warmth level fitted separately
pd <- ggplot2::position_dodge(width = 0.4)
## quarterly plot setting starts -------------------------
energy_type = ""
## energy_type = "_gas"
duration = "quarterly"
additive_keywords = "nonsingle_"
if (additive_keywords == "") {
df_recode_retrofit = data.frame(header=sprintf("X%s", 2:7),
retrofit=c("Advanced Metering", "Bundle", "Building Envelope", "GSALink", "HVAC",
"Lighting"))
} else if (additive_keywords == "nonsingle_") {
df_recode_retrofit = data.frame(header=sprintf("X%s", 2:7),
retrofit=c("Advanced Metering", "Building Envelope", "Commissioning", "GSALink",
"HVAC", "Lighting"))
}
endbound = 20
fatten_amount = 2.5
size = 0.2
point_size = 1.0
## quarterly plot setting starts -------------------------
## monthly plot setting starts -------------------------
## energy_type = ""
## ## energy_type = "_gas"
## duration = "monthly"
## endbound = 60
## fatten_amount = 2.5
## size = 0.2
## point_size = 1.0
## produce plots with all 4 model type
## models = c("no_control", "hddcdd", "hdd2cdd2", "bin")
## produce plots with only one model type
models = c("bin")
## monthly plot setting ends -------------------------
## yearly plot setting starts -------------------------
## duration = "yearly"
## endbound = 5
## fatten_amount = 2.5
## size = 1.0
## point_size = 2.5
## yearly plot setting ends -------------------------
acc <- lapply(models, function(model_type) {
dfacc =
readr::read_tsv(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/%s_%sretrofit_no_climate_%s%s.txt", duration, additive_keywords, model_type, energy_type), skip=7, col_names=FALSE) %>%
## remove bottom non-conformative rows
head(-5) %>%
dplyr::mutate(`variable`=rep(c("coef", "se", "ci_low", "ci_high"), nrow(.)/4)) %>%
tidyr::fill(`X1`) %>%
tidyr::unite(`temp`, `X1`, `variable`, sep="-") %>%
## modify this to X2:X?, ? is the number of models + 1
tidyr::gather(`header`, `value`, X2:X7) %>%
tidyr::separate(`temp`, c("X1", "variable"), sep="-") %>%
dplyr::filter(`variable` != "se",
!grepl(".eDate", `X1`, fixed = TRUE),
!grepl("cdd", `X1`, fixed=TRUE),
!grepl("hdd", `X1`, fixed=TRUE),
!grepl("Constant", `X1`, fixed=TRUE),
!grepl("avgTemp", `X1`, fixed=TRUE),
) %>%
dplyr::mutate(`value`=gsub("\\(", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\)", "", `value`)) %>%
dplyr::mutate(`value`=gsub("\\*", "", `value`)) %>%
dplyr::mutate(`value`=gsub(",", "", `value`)) %>%
tidyr::unite(`temp`, `X1`, `header`, sep="-") %>%
tidyr::spread(variable, value) %>%
tidyr::separate(`temp`, c("X1", "header"), sep="-") %>%
dplyr::filter(!is.na(`coef`)) %>%
dplyr::mutate(`coef`=as.numeric(`coef`),
`ci_high`=as.numeric(`ci_high`),
`ci_low`=as.numeric(`ci_low`),
) %>%
dplyr::mutate(`tau`=as.integer(substr(X1, 1, stringi::stri_locate(X1, fixed=".", mode="first")-1))) %>%
dplyr::mutate(`tau`=ifelse(`tau` <= endbound, `tau`, `tau` - 2*endbound - 1)) %>%
dplyr::left_join(df_recode_retrofit, by="header") %>%
dplyr::mutate(`model`=model_type) %>%
dplyr::select(-`X1`, -`header`) %>%
{.}
})
modelResult = do.call(rbind, acc)
unique(modelResult$model)
tail(modelResult)
if (additive_keywords == "") {
retrofits = c("Advanced Metering", "Building Envelope", "HVAC", "Lighting", "GSALink", "Bundle")
names(retrofits) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
ylower = c(-1, -1, -4.5, -1.5, -5, -1)
names(ylower) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
yupper = c(1.2, 0.9, 2.2, 1.4, 5, 0.6)
names(yupper) = c("metering", "envelope", "hvac", "lighting", "GSALink", "bundle")
} else if (additive_keywords == "nonsingle_") {
retrofits = c("Advanced Metering", "Building Envelope", "HVAC", "Lighting", "GSALink", "Commissioning")
names(retrofits) = c("metering", "envelope", "hvac", "lighting", "GSALink", "commissioning")
ylower = c(-0.6, -0.8, -0.6, -0.75, -2, -0.75)
names(ylower) = c("metering", "envelope", "hvac", "lighting", "GSALink", "commissioning")
yupper = c(0.2, 0.25, 0.2, 0.25, 1.5, 0.25)
names(yupper) = c("metering", "envelope", "hvac", "lighting", "GSALink", "commissioning")
}
if (energy_type == "") {
energy_label = "Electricity"
} else if (energy_type == "_gas") {
energy_label = "Gas"
}
slide_size = 22
for (fileSuffix in names(retrofits)) {
retrofit_type <- retrofits[[fileSuffix]]
print(retrofit_type)
p <-
modelResult %>%
dplyr::filter(retrofit == retrofit_type) %>%
dplyr::bind_rows(data.frame(`tau`=0, `ci_low`=0, `ci_high`=0, `coef`=0, `retrofit`=retrofit_type, `model`=models)) %>%
ggplot2::ggplot(ggplot2::aes(x=tau, y=coef)) +
ggplot2::geom_line() +
ggplot2::geom_pointrange(ggplot2::aes(x=tau, y=`coef`, ymin=`ci_low`, ymax=`ci_high`), position=pd, shape=18, fatten=fatten_amount, size=size) +
ggplot2::geom_point(x=0, y=0, shape=18, size=point_size)
if (length(models) == 1) {
p <- p +
ggplot2::facet_wrap(`model`~.)
} else {
p <- p +
ggplot2::facet_wrap(`model`~., ncol=2)
}
p <- p +
ggplot2::xlab("Number of periods before or after retrofit") +
ggplot2::ylab(NULL) +
ggplot2::theme_bw() +
ggtitle(sprintf("%s Event study model of \"%s\"", energy_label, retrofit_type))
if (!is.null(ylower)) {
p <- p +
ggplot2::coord_cartesian(ylim=c(ylower[[fileSuffix]], yupper[[fileSuffix]]))
}
p <- p +
ggplot2::theme(legend.position = "bottom",
plot.title = element_text(size=slide_size))
print(p)
ggplot2::ggsave(file=sprintf("~/Dropbox/thesis/code/pubPriCmp/image/%s_%sretrofit_no_climate_coef_ci_%s_%s%s.png", duration, additive_keywords, fileSuffix, length(models), energy_type), width=8, height=5, units="in")
}
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
## start section: create summary of energy, weather, and building count for 4 alternative study set, based on different building types
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
get.study.set <- function(own.or.lease, office.class) {
if (!is.na(office.class)) {
if (office.class == "o_o") {
generic.office.pnc = "Office"
generic.office.gsa = "Office"
} else if (office.class == "o_oct") {
generic.office.pnc = "Office"
generic.office.gsa = c("Courthouse", "Office", "CT/Office")
} else if (office.class == "r_o") {
generic.office.pnc = "Retail"
generic.office.gsa = "Office"
} else if (office.class == "ro_oct") {
generic.office.pnc = c("Retail", "Office")
generic.office.gsa = c("Courthouse", "Office", "CT/Office")
}
to.summarize = allDataGreen %>%
dplyr::filter((Organization == "PNC" & `type_general` %in% generic.office.pnc) | (Organization == "GSA" & `type_general` %in% generic.office.gsa)) %>%
{.}
} else {
to.summarize = allDataGreen
generic.office.gsa = NA
generic.office.pnc = NA
}
if (!is.na(own.or.lease)) {
to.summarize <- to.summarize %>%
dplyr::filter(`Ownership`==own.or.lease) %>%
{.}
}
return(list(df=to.summarize, generic.office.gsa=generic.office.gsa, generic.office.pnc=generic.office.pnc))
}
## following is how to use the above function
## restrict to owned public offices, courthouse, private offices
## temp <- get.study.set("Owned", "o_oct")
## restrict to owned and leased public offices, courthouse, private offices
## temp <- get.study.set(NA, "o_oct")
## No restrictions
## temp <- get.study.set(NA, NA)
temp$df
## following check on what is returned by the above function
temp %>%
distinct(Name, `type_general`, Organization) %>%
dplyr::group_by(Organization, `type_general`) %>%
dplyr::summarise(n()) %>%
print()
folder.suffixes = c("",
"_o_o_owned", "_o_oct_owned", "_r_o_owned", "_ro_oct_owned",
"_o_o_owned_and_leased", "_o_oct_owned_and_leased", "_r_o_owned_and_leased", "_ro_oct_owned_and_leased"
)
captions = c("no restrictions",
"Public: owned offices; Private: owned offices",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices",
"Public: owned offices; Private: owned retails",
"Public: owned offices, courthouses, CT/Offices; Private: owned offices, retails",
"Public: offices; Private: offices",
"Public: offices, courthouses, CT/Offices; Private: offices",
"Public: offices; Private: retails",
"Public: offices, courthouses, CT/Offices; Private: offices, retails")
## change the major source of greeness to howe by county
data.to.summarize = allDataGreen %>%
dplyr::mutate(Greeness = greeness.howe.county) %>%
{.}
for (i in seq_along(folder.suffixes)) {
s = folder.suffixes[i]
print(s)
tokens = unlist(strsplit(s, "_"))
if (length(tokens) == 0) {
ownlease = NA
usetype = NA
} else {
usetype = paste0(tokens[2:3], collapse = "_")
if ("leased" %in% tokens) {
ownlease = NA
} else {
ownlease = "Owned"
}
}
if (is.na(ownlease)) {
own.or.lease.str = ""
} else {
own.or.lease.str = ownlease
}
result <- get.study.set(ownlease, usetype)
to.summarize = result$df
generic.office.gsa = result$generic.office.gsa
generic.office.pnc = result$generic.office.pnc
## may not be the properway to check whether the vector is NA
if (is.na(generic.office.gsa[1])) {
generic.office.gsa = "no restriction"
}
if (is.na(generic.office.pnc[1])) {
generic.office.pnc = "no restriction"
}
to.summarize %>%
distinct(Name, `type_general`, Organization) %>%
dplyr::group_by(Organization, `type_general`) %>%
dplyr::summarise(n()) %>%
print()
df.n = to.summarize %>%
dplyr::distinct(Organization, Name) %>%
dplyr::group_by(Organization) %>%
dplyr::summarise(cnt=n()) %>%
dplyr::ungroup() %>%
tibble::column_to_rownames("Organization") %>%
{.}
## slide summary stats
df1 = to.summarize %>%
dplyr::mutate(`<40`=`<10` + `[10-20)` + `[20-30)` + `[30-40)`,
`>80`=`[80-90)` + `>90`) %>%
dplyr::select(`eui_elec`, `eui_gas`, `<40`, `>80`, `private`, `Name`) %>%
dplyr::group_by(`private`) %>%
dplyr::summarise(`eui_elec`=mean(`eui_elec`), `eui_gas`=mean(`eui_gas`), `<40`=mean(`<40`), `>80`=mean(`>80`), `building count`=n_distinct(Name)) %>%
dplyr::ungroup() %>%
{.}
df2 = to.summarize %>%
dplyr::mutate(`<40`=`<10` + `[10-20)` + `[20-30)` + `[30-40)`,
`>80`=`[80-90)` + `>90`) %>%
dplyr::select(`eui_elec`, `eui_gas`, `<40`, `>80`, `private`, `region_wiki`, `Name`) %>%
dplyr::group_by(`private`, `region_wiki`) %>%
dplyr::summarise(`eui_elec`=mean(`eui_elec`), `eui_gas`=mean(`eui_gas`), `<40`=mean(`<40`), `>80`=mean(`>80`), `building count`=n_distinct(Name)) %>%
dplyr::ungroup() %>%
{.}
df3 = to.summarize %>%
dplyr::mutate(`<40`=`<10` + `[10-20)` + `[20-30)` + `[30-40)`,
`>80`=`[80-90)` + `>90`) %>%
dplyr::select(`eui_elec`, `eui_gas`, `<40`, `>80`, `private`, `Greeness`, `Name`) %>%
dplyr::group_by(`private`, `Greeness`) %>%
dplyr::summarise(`eui_elec`=mean(`eui_elec`), `eui_gas`=mean(`eui_gas`), `<40`=mean(`<40`), `>80`=mean(`>80`), `building count`=n_distinct(Name)) %>%
dplyr::ungroup() %>%
{.}
sink(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/energy_weather_n_by_region_green%s.tex",
s))
df1 %>%
dplyr::bind_rows(df2, df3) %>%
dplyr::select(private, `region_wiki`, Greeness, everything()) %>%
dplyr::mutate(order = c(1, 6, 2, 3, 7, 8, 4, 5, 9, 10)) %>%
dplyr::arrange(order) %>%
## dplyr::mutate(private = c("public", rep("", 5), "private", rep("", 5))) %>%
dplyr::mutate(private=ifelse(private == 1, "private", "public")) %>%
dplyr::mutate(` `=ifelse(is.na(`region_wiki`), ifelse(is.na(`Greeness`), private, `Greeness`), `region_wiki`)) %>%
dplyr::select(-`region_wiki`, -`Greeness`, -private, -order) %>%
dplyr::select(` `, everything()) %>%
dplyr::rename(`Electricity`=`eui_elec`, `Gas`=`eui_gas`) %>%
knitr::kable("latex", booktabs = T, format.args=list(big.mark=",", digits=3),
caption=sprintf("%s PNC (n=%s): %s; GSA (n=%s): %s",
own.or.lease.str,
df.n["PNC",1],
paste(generic.office.pnc, collapse = ", "),
df.n["GSA",1],
paste(generic.office.gsa, collapse = ", "))) %>%
kableExtra::kable_styling(full_width = FALSE, font_size = 7) %>%
kableExtra::group_rows("US region", 2, 3, bold=FALSE) %>%
kableExtra::group_rows("Greeness", 4, 5, bold=FALSE) %>%
kableExtra::group_rows("US region", 7, 8, bold=FALSE) %>%
kableExtra::group_rows("Greeness", 9, 10, bold=FALSE) %>%
kableExtra::column_spec(2:6, width = "1cm") %>%
## kableExtra::kable_styling(latex_options = "striped") %>%
kableExtra::add_header_above(c(" "=1, "Monthly consumption\n(kBtu/sqft)" = 2, "# of days in a month \nwith mean temperature" = 2, "n" = 1), escape = TRUE) %>%
print()
sink()
}
allDataGreen %>%
dplyr::distinct(`Name`, `Organization`, `region_wiki`) %>%
dplyr::group_by(`Organization`, `region_wiki`) %>%
dplyr::summarise(n())
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
## end section: create summary of energy, weather, and building count for 4 alternative study set, based on different building types
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
## comparing static characteristics / balance of covariates of the two portfolio
load("../data/allDataGreen.rda")
balance.check.vars = c("GSF", "Sum of Employees", "Electric_(kBtu)",
"Gas_(kBtu)","eui_elec", "eui_gas", "eui_total", "HDD",
"CDD", "<10", "[20-30)", "[30-40)", "[40-50)" ,
"[50-60)", "[60-70)" , "[70-80)", "[80-90)" , ">90")
for (i in seq_along(folder.suffixes)) {
s = folder.suffixes[i]
print(s)
tokens = unlist(strsplit(s, "_"))
if (length(tokens) == 0) {
ownlease = NA
usetype = NA
} else {
usetype = paste0(tokens[2:3], collapse = "_")
if ("leased" %in% tokens) {
ownlease = NA
} else {
ownlease = "Owned"
}
}
if (is.na(ownlease)) {
own.or.lease.str = ""
} else {
own.or.lease.str = ownlease
}
result <- get.study.set(ownlease, usetype)
studyset = result$df
generic.office.gsa = result$generic.office.gsa
generic.office.pnc = result$generic.office.pnc
## may not be the properway to check whether the vector is NA
if (is.na(generic.office.gsa[1])) {
generic.office.gsa = "no restriction"
}
if (is.na(generic.office.pnc[1])) {
generic.office.pnc = "no restriction"
}
studyset %>%
distinct(Name, `type_general`, Organization) %>%
dplyr::group_by(Organization, `type_general`) %>%
dplyr::summarise(n()) %>%
print()
df.n = studyset %>%
dplyr::distinct(Organization, Name) %>%
dplyr::group_by(Organization) %>%
dplyr::summarise(cnt=n()) %>%
dplyr::ungroup() %>%
tibble::column_to_rownames("Organization") %>%
{.}
acc.balance = NULL
for (v in balance.check.vars) {
tresult <- t.test(as.formula(sprintf("`%s` ~ private", v)), data = studyset, alternative = "two.sided")
mu0 <- tresult$estimate[[1]]
mu1 <- tresult$estimate[[2]]
dif <- mu1 - mu0
p <- tresult$p.value
p.with.sig <- ifelse(p < 0.05, ifelse(p >= 0.01, sprintf("%.2f*", p), sprintf("%.2f**", p)), sprintf("%.2f", p))
print(p)
print(p.with.sig)
acc.balance <- rbind(acc.balance,
data.frame(variable = v, mu0 = mu0, mu1 = mu1, dif = dif, p = p, p.with.sig = p.with.sig))
}
sink(sprintf("~/Dropbox/thesis/writeups/policy_cmp/tables/balance_compare%s.tex", s))
acc.balance %>%
dplyr::select(-p) %>%
knitr::kable("latex", booktabs = T,
format.args=list(big.mark=",", digits=2, scientific=F),
caption=sprintf("%s PNC (n=%s): %s; GSA (n=%s): %s",
own.or.lease.str,
df.n["PNC",1],
paste(generic.office.pnc, collapse = ", "),
df.n["GSA",1],
paste(generic.office.gsa, collapse = ", "))) %>%
kableExtra::kable_styling(full_width = FALSE, font_size = 7) %>%
kableExtra::group_rows("Static feature", 1, 2, bold=FALSE) %>%
kableExtra::group_rows("Energy", 3, 7, bold=FALSE) %>%
kableExtra::group_rows("Weather", 8, 18, bold=FALSE) %>%
## kableExtra::column_spec(2:6, width = "1cm") %>%
print()
sink()
}
dfstatic = allDataGreen %>%
dplyr::select(-`latitude`, -`longitude`, -`City`, -`State`, -`year`, -`Organization`, -`Date`) %>%
dplyr::mutate(Portfolio=ifelse(private==0, "public", "private")) %>%
{.}
## comparing building types of the two portfolio
num.building.per.type <- dfstatic %>%
dplyr::distinct(Name, Portfolio, `type_general`) %>%
dplyr::group_by(Portfolio, `type_general`) %>%
dplyr::summarise(n()) %>%
dplyr::ungroup() %>%
tidyr::spread(`Portfolio`, `n()`, fill=0L) %>%
dplyr::mutate(ordering = pmax(private, public)) %>%
dplyr::arrange(desc(ordering)) %>%
dplyr::select(-ordering) %>%
dplyr::mutate_at(vars(`type_general`), ~replace(., is.na(.), "Unknown")) %>%
{.}
result_table_tex <- xtable(num.building.per.type, caption="Building type count")
print(result_table_tex, tabular.environment = "tabular", include.rownames=FALSE,
file="~/Dropbox/thesis/writeups/policy_cmp/tables/num_building_per_type.tex")
## comparing static info of the two portfolio
s1 =
dfstatic %>%
dplyr::group_by(Portfolio) %>%
dplyr::summarise(GSF=mean(GSF), `Sum of Employees`=mean(`Sum of Employees`)) %>%
dplyr::ungroup() %>%
tibble::column_to_rownames("Portfolio") %>%
t() %>%
as.data.frame() %>%
tibble::rownames_to_column("variable") %>%
{.}
s2 =
dfstatic %>%
dplyr::mutate(Ownership = sprintf("%s percent", Ownership)) %>%
dplyr::group_by(Portfolio, Ownership) %>%
dplyr::summarise(count=n()) %>%
dplyr::ungroup() %>%
dplyr::group_by(Portfolio) %>%
dplyr::mutate(percentage=count/sum(count) * 100) %>%
dplyr::ungroup() %>%
dplyr::select(-count) %>%
tidyr::spread(Portfolio, `percentage`) %>%
dplyr::rename(variable=Ownership) %>%
{.}
s3 =
dfstatic %>%
dplyr::mutate(`Building Type`=ifelse(`Building Type` %in% c("Office", "Retail"), `Building Type`, "Other")) %>%
dplyr::group_by(Portfolio, `Building Type`) %>%
dplyr::summarise(count=n()) %>%
dplyr::ungroup() %>%
dplyr::group_by(Portfolio) %>%
dplyr::mutate(percentage = count/sum(count) * 100) %>%
dplyr::ungroup() %>%
dplyr::select(-count) %>%
tidyr::spread(Portfolio, `percentage`, fill=0) %>%
dplyr::rename(variable=`Building Type`) %>%
dplyr::arrange(desc(private)) %>%
{.}
t.test(dfstatic %>% dplyr::filter(Portfolio == "private") %>% .$GSF,
dfstatic %>% dplyr::filter(Portfolio == "public") %>% .$GSF)
static.cmp =
s1 %>%
bind_rows(s2) %>%
bind_rows(s3) %>%
dplyr::mutate(category=c("size", "", "Ownership", "", "", "Type", "", "")) %>%
dplyr::select(category, everything()) %>%
{.}
result_table_tex <- xtable(static.cmp, caption="Building static feature comparison")
print(result_table_tex, tabular.environment = "tabular", include.rownames=FALSE,
file="~/Dropbox/thesis/writeups/policy_cmp/tables/static_feature_cmp.tex")
allData %>%
dplyr::distinct(Organization, State, Name) %>%
dplyr::group_by(Organization, State) %>%
print()
## Plot GSA energy along with the policy change timeline
devtools::load_all("~/Dropbox/gsa_2017/db.interface")
gsa_energy =
db.interface::read_table_from_db(dbname = "all", tablename = "EUAS_monthly",
cols=c("Building_Number", "state_abbr", "Gross_Sq.Ft", "year", "month",
"Electric_(kBtu)", "Gas_(kBtu)", "Gross_Sq.Ft"
##, "Occupancy", "datacenter_sqft", "lab_sqft"
)) %>%
dplyr::rename(`Name`=`Building_Number`) %>%
## na.omit() %>%
{.}
## plot energy
toplot = gsa_energy %>%
dplyr::rename(`GSF`=`Gross_Sq.Ft`) %>%
dplyr::filter(`GSF`>0) %>%
dplyr::mutate(`Energy`=`Electric_(kBtu)` + `Gas_(kBtu)`) %>%
dplyr::group_by(year, month) %>%
dplyr::summarise(Energy=sum(Energy), GSF=sum(GSF)) %>%
dplyr::ungroup() %>%
## dplyr::mutate(target=Energy) %>%
dplyr::mutate(target=Energy/GSF) %>%
{.}
ylimit = max(toplot$target)
toplot %>%
dplyr::mutate(Time=zoo::as.yearmon(sprintf("%d-%d", year, month), format="%Y-%m")) %>%
ggplot2::ggplot(ggplot2::aes(x=Time, y=target)) +
ggplot2::geom_line() +
ggplot2::geom_vline(xintercept=as.yearmon("2005-08", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2005-08", format="%Y-%m"), y=1.0*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2005-08", format="%Y-%m"), y=1.0*ylimit, label="EPAct 2005") +
ggplot2::geom_vline(xintercept=as.yearmon("2006-01", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2006-01", format="%Y-%m"), y=0.9*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2006-01", format="%Y-%m"), y=0.9*ylimit, label="Guiding Principles\n(Memorandum in 2006)") +
ggplot2::geom_vline(xintercept=as.yearmon("2007-01", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2007-01", format="%Y-%m"), y=0.95*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2007-01", format="%Y-%m"), y=0.95*ylimit, label="Executive Order 13423") +
ggplot2::geom_vline(xintercept=as.yearmon("2007-12", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2007-12", format="%Y-%m"), y=1.0*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2007-12", format="%Y-%m"), y=1.0*ylimit, label="EISA 2007") +
ggplot2::geom_vline(xintercept=as.yearmon("2008-12", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2008-12", format="%Y-%m"), y=0.95*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2008-12", format="%Y-%m"), y=0.95*ylimit, label="Guiding Principles\n(revision in 2008)") +
ggplot2::geom_vline(xintercept=as.yearmon("2009-02", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2009-02", format="%Y-%m"), y=0.9*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2009-02", format="%Y-%m"), y=0.9*ylimit, label="Recovery Act") +
ggplot2::geom_vline(xintercept=as.yearmon("2009-10", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2009-10", format="%Y-%m"), y=1.0*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2009-10", format="%Y-%m"), y=1.0*ylimit, label="Executive Order 13514") +
ggplot2::geom_vline(xintercept=as.yearmon("2011-02", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2011-02", format="%Y-%m"), y=0.9*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2011-02", format="%Y-%m"), y=0.9*ylimit, label="President Memorandum") +
ggplot2::geom_vline(xintercept=as.yearmon("2011-06", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2011-06", format="%Y-%m"), y=0.95*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2011-06", format="%Y-%m"), y=0.95*ylimit, label="Better Buildings Initiative") +
ggplot2::geom_vline(xintercept=as.yearmon("2015-03", format="%Y-%m"), linetype="dashed") +
geom_point(aes(x=as.yearmon("2015-03", format="%Y-%m"), y=0.9*ylimit), colour="red") +
ggplot2::annotate("text", x=as.yearmon("2015-03", format="%Y-%m"), y=0.9*ylimit, label="Excecutive Order 13693") +
ggplot2::theme_bw() +
ggplot2::ylab("Electricity + Gas (kBtu/sqft)") +
ggplot2::ggtitle("The energy consumption trend and policy timelines of the public portfolio") +
ggplot2::theme()
ggplot2::ggsave(file="~/Dropbox/thesis/code/pubPriCmp/image/gsa_eui_trend_policy_time.png", width=13, height=5, units="in")
## plot GSF
ylimit = max(toplot$GSF)
toplot %>%
dplyr::mutate(Time=zoo::as.yearmon(sprintf("%d-%d", year, month), format="%Y-%m")) %>%
ggplot2::ggplot(ggplot2::aes(x=Time, y=GSF)) +
ggplot2::geom_line() +
ggplot2::geom_vline(xintercept=as.yearmon("2010-06", format="%Y-%m"), linetype="dashed") +
ggplot2::annotate("text", x=as.yearmon("2010-06", format="%Y-%m"), y=0.9*ylimit, label="Memorandum June 2010\n(Disposing of Unneeded Federal Real Estate)") +
ggplot2::theme_bw() +
ggplot2::ggtitle("The GSF and time line the public portfolio") +
ggplot2::theme()
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.