R/get-acs-recs.R
In psrc/psrccensus: Summarize Census Data for use at PSRC
Defines functions
add_acs_share
get_acs_recs
reliability_calcs
label_acs_variables
acs_varsearch
Documented in
acs_varsearch
add_acs_share
get_acs_recs
label_acs_variables
reliability_calcs
#' @importFrom magrittr %<>% %>%
NULL
#' Search published ACS variables
#'
#' Identify desired tables by examining prefix of relevant variable codes
#' Includes primary, subject and profile tables
#' @param regex search term
#' @param year optionally restrict search to a specific year
#' @return data.table of filtered variable codes and attributes
#' @author Michael Jensen
#'
#' @rawNamespace import(data.table, except = c(month, year))
#' @importFrom lubridate now month year
#' @export
acs_varsearch <- function(regex, year=NULL){
acs_year <- if(is.null(year)){year(now() - months(18))}else{year}
name <- label <- concept <- NULL # Instantiate tidycensus::pums_variables variable locally (for documentation, not function)
pull_varlist <- function(survey){
x <- tidycensus::load_variables(acs_year, survey) %>% setDT() %>%
.[grepl(regex, name, ignore.case=TRUE)|
grepl(regex, label, ignore.case=TRUE)|
grepl(regex, concept, ignore.case=TRUE)] %>% unique()
return(x)
}
acstypes <- paste0("acs5", c("","/subject","/profile","/cprofile")) %>% c("acsse", recursive=TRUE)
rs <- list()
rs <- lapply(acstypes, pull_varlist) %>% rbindlist(fill=TRUE)
return(rs)
}
#' Label ACS variables
#'
#' Helper function to provide variable labels and concept--i.e. topic--alongside codes
#' @param df dataframe with Census API result
#' @param table.name Census table code
#' @param year the year--or last year--of the ACS survey
#' @param acs.type either acs1 or acs5
#'
#' @return dataframe with labels appended
label_acs_variables <- function(df, table.name, year, acs.type){
name <- label <- concept <- NULL # instantiate variables
survey <- if(table.name %in% acs_tbltypes_lookup$subject){paste0(acs.type,"/subject")
}else if(table.name %in% acs_tbltypes_lookup$profile){paste0(acs.type,"/profile")
}else if(table.name %in% acs_tbltypes_lookup$cprofile){paste0(acs.type,"/cprofile")
}else if(table.name %in% acs_tbltypes_lookup$acsse){"acsse"
}else{acs.type}
labels <- tidycensus::load_variables(year=year, dataset=survey) %>%
dplyr::select(c(name, label, concept)) %>% dplyr::rename(variable=name)
df <- dplyr::left_join(df, labels, by=c("variable"))
return(df)
}
#' ACS Estimates by County
#'
#' Generate ACS estimates for multiple tables by multiple counties
#' This is a helper function for the main function get_acs_recs.
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param counties A character string or vector of counties. Defaults to PSRC counties.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#' @param acs.type A character string as either 'acs1', 'acs3' or acs5'.
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by counties for selected table codes and years with a regional aggregation.
#' Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@importFrom dplyr filter
get_acs_county <- function (state="Washington", counties = c("King","Kitsap","Pierce","Snohomish"), table.names, years, acs.type) {
census.data <- geography <- NAME <- name <- variable <- NULL
estimate <- moe <- sumest <- summoe <- label <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(state=state, geography='county', county=counties, year=year, survey=acs.type, table=table)
# Split County Name for County and State
tbl <- tbl %>% tidyr::separate(col=NAME, into=c("name", "state"),sep=",") %>%
dplyr::mutate(estimate =tidyr::replace_na(estimate,0))
tbl$state <- trimws(tbl$state, "l")
# Create a Total for Region if all 4 PSRC Counties are pulled and then join to county table
if (identical(counties, c("King","Kitsap","Pierce","Snohomish"))){
total <- tbl %>%
dplyr::select(variable, estimate, moe) %>%
dplyr::mutate(estimate =tidyr::replace_na(estimate,0))%>%
dplyr::group_by(variable) %>%
dplyr::summarize(sumest = sum(estimate), summoe = tidycensus::moe_sum(moe, estimate)) %>%
dplyr::rename(estimate=sumest, moe=summoe) %>%
dplyr::mutate(GEOID="REGION", name="Region", state=state)
tbl <- dplyr::bind_rows(list(tbl,total))
}
# Add labels, column for Census Geography, Type and Year of Data
tbl <- tbl %>% label_acs_variables(table, year, acs.type) %>%
dplyr::mutate(census_geography="County", acs_type = acs.type, year=year)
# Median and average calculations are more complicated, may need PUMS, filter out for now:
tbl <- tbl %>% dplyr::filter(name !='Region' |
(!(stringr::str_detect(label, 'Median'))&!(stringr::str_detect(label, 'Average'))))
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by MSA
#'
#' Generate ACS estimates for multiple tables by multiple MSA's
#'
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#' @param acs.type A character string as either 'acs1', 'acs3' or acs5'.
#' @param FIPS Character string for FIPS codes for specific MSA geographies. Defaults to Seattle & Bremerton MSA c("14740","42660")
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by MSA for selected table codes and years. Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@keywords internal
get_acs_msa <- function (table.names, years, acs.type, FIPS = c("14740","42660")) {
census.data <- geography <- NAME <- name <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(geography="metropolitan statistical area/micropolitan statistical area", year=year, survey=acs.type, table=table) %>%
dplyr::filter(GEOID %in% FIPS) %>%
tidyr::separate(col=NAME, into=c("name", "state"),sep=",")
tbl$state <- trimws(tbl$state, "l")
# Add labels, column for Census Geography, Type and Year of Data
tbl %<>% label_acs_variables(table, year, acs.type) %>%
dplyr::mutate(census_geography="MSA", acs_type=acs.type, year=year)
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by Place
#'
#' Generate ACS estimates for multiple tables by multiple places
#'
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#' @param acs.type A character string as either 'acs1', 'acs3' or acs5'.
#' @param place_FIPS Character string of FIPS codes (with state prefix) for specific Census Places. If NULL, Places within the PSRC Region will be returned.
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by Place for selected table codes and years. Includes detailed variable names.
#'
#' @importFrom rlang .data
#' @keywords internal
get_acs_place <- function (state="Washington", table.names, years, acs.type, place_FIPS=NULL) {
census.data <- geography <- NAME <- name <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Determine Places within Region
if(is.null(place_FIPS)){psrc_places <- get_psrc_places(year) %>% sf::st_drop_geometry() %>% unique() %>%
dplyr::pull(GEOID)}
# Download ACS Data
tbl <- tidycensus::get_acs(state=state, geography='place', year=year, survey=acs.type, table=table) %>%
tidyr::separate(col=NAME, into=c("name", "state"),sep=",")
if(!is.null(place_FIPS)){tbl %<>% filter(GEOID %in% place_FIPS)}else{tbl %<>% filter(GEOID %in% psrc_places)}
tbl$state <- trimws(tbl$state, "l")
# Add labels, column for Census Geography, Type and Year of Data
tbl %<>% label_acs_variables(table, year, acs.type) %>%
dplyr::mutate(acs_type = acs.type, year=year) %>%
dplyr::mutate(census_geography = dplyr::case_when(
endsWith(name, "city") ~ "City",
endsWith(name, "CDP") ~ "CDP")) %>%
dplyr::mutate(name = gsub(" city", "", name)) %>%
dplyr::mutate(name = gsub(" CDP", "", name))
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by Census Tract
#'
#' Generate ACS 5yr estimates for multiple tables by Census Tracts in multiple counties
#'
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param counties A character string or vector of counties. Defaults to PSRC counties.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by tracts for selected table codes. Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@keywords internal
get_acs_tract <- function (state="Washington", counties = c("King","Kitsap","Pierce","Snohomish"), table.names, years) {
census.data <- geography <- NAME <- name <- county <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(state=state, county=counties, geography='tract', year=year, survey="acs5", table=table) %>%
tidyr::separate(col=NAME, into=c("name", "county", "state"),sep=",")
tbl$county <- trimws(tbl$county, "l")
tbl$state <- trimws(tbl$state, "l")
county.names <- paste(counties,"County")
# Add label, column for Census Geography, Type and Year of Data
tbl %<>% label_acs_variables(table, year, "acs5") %>%
dplyr::mutate(census_geography="Tract", acs_type = "acs5", year=year) %>%
dplyr::select(-county)
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by Census Block Group
#'
#' Generate ACS 5yr estimates for multiple tables by Census Block Groups in multiple counties
#'
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param counties A character string or vector of counties. Defaults to PSRC counties.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by tracts for selected table codes. Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@keywords internal
get_acs_blockgroup <- function (state="Washington", counties = c("King","Kitsap","Pierce","Snohomish"), table.names, years) {
census.data <- geography <- NAME <- name <- county <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(state=state, county=counties, geography="block group", year=year, survey="acs5", table=table) %>%
tidyr::separate(col=NAME, into=c("name", "tract", "county", "state"),sep=",")
tbl$county <- trimws(tbl$county, "l")
tbl$state <- trimws(tbl$state, "l")
county.names <- paste(counties,"County")
# Add labels, column for Census Geography, Type and Year of Data
tbl %<>% label_acs_variables(table, year, "acs5") %>%
dplyr::mutate(census_geography="block group", acs_type="acs5", year=year) %>%
dplyr::select(-county)
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by Census PUMA
#'
#' Generate ACS 5 year estimates for multiple tables by Census PUMA
#'
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#'
#' @author Christy Lam
#'
#' @return a tibble of acs estimates by PUMA for selected table codes. Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@keywords internal
get_acs_puma <- function (state = "Washington", table.names, years) {
census.data <- geography <- NAME <- name <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(state = state, geography = "public use microdata area", year = year, table = table)
# separate state from PUMA name, clean-up name column
tbl <- tbl %>%
tidyr::separate(col = NAME, into = c("name", "state"), sep = "Washington") %>%
dplyr::mutate(state = 'Washington') %>%
dplyr::mutate(name = stringr::str_replace_all(name, "[\\;|\\,]\\s*$", ""))
# Add labels, column for Census Geography, Type and Year of Data
tbl <- tbl %>%
label_acs_variables(table, year, "acs5") %>%
dplyr::mutate(census_geography = "public use microdata area (PUMA)", acs_type = "acs5", year = year)
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by State
#'
#' Generate ACS estimates for multiple tables by state
#' This is a helper function for the main function get_acs_recs.
#' @param states A character string of state names or abbreviations. Defaults to Washington.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#' @param acs.type A character string as either 'acs1', 'acs3' or acs5'.
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by states for selected table codes.
#' Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@importFrom dplyr filter
get_acs_state <- function (states="Washington", table.names, years, acs.type) {
census.data <- geography <- NAME <- name <- variable <- NULL
estimate <- moe <- sumest <- summoe <- label <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(geography="state", state=states, year=year, survey=acs.type, table=table) %>%
dplyr::mutate(estimate =tidyr::replace_na(estimate,0))
# Format fields for compatibility
tbl <- tbl %>%
tidyr::separate(col = NAME, into = c("name", "state"), sep = "Washington") %>%
dplyr::mutate(state = "Washington", name = "Washington State")
# Add labels, column for Census Geography, Type and Year of Data
tbl <- tbl %>% label_acs_variables(table, year, acs.type) %>%
tidyr::replace_na(list(moe=0)) %>%
dplyr::mutate(census_geography = "state", acs_type = acs.type, year = year)
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' Add estimate reliability information
#'
#' After gathering data, add reliability information using moe and estimate
#'
#' @param dfs data frames already retrieved from the api, or created by a user
#' @param moe character name of the column that has contains margin of error estimates
#' @param estimate character name of the column that has contains data value estimates
#' @author Suzanne
#' @return the data frames with new columns for se, cv, and reliability
#' @export
reliability_calcs<- function(dfs, moe='moe', estimate='estimate'){
# A coefficient of variation (CV) measures the relative amount of sampling error that is associated with a sample
# estimate. The CV is calculated as the ratio of the SE for an estimate to the estimate itself and is usually
# expressed as a percent.
# Note that since both the ACS and household travel survey are reported using a 90 percent confidence interval,
# where the Margin of Error (MOE) is reported in place of standard error, you can convert it to standard error by
# dividing by 1.645.
# to do: put 1.645 in a better place in the package, it's a magic number
# http://aws-linux/mediawiki/index.php/Understanding_Error_and_Determining_Statistical_Significance
se <- NULL # Bind variables locally (for documentation, not function)
zscore_90<-1.645 # Conversion factor from standard error to 90 percent confidence interval
dfs%<>%dplyr::mutate(se=!!rlang::ensym(moe)/zscore_90)%>%
dplyr::mutate(cv= se/!!rlang::ensym(estimate))%>%
dplyr::mutate(reliability=
dplyr::case_when(
!!rlang::ensym(estimate)==0 ~ 'estimate is 0, cannot compute',
cv<=0.15 ~ 'good',
cv>0.15 & cv<=0.30 ~ 'fair',
cv>0.30 & cv<=0.50 ~ 'use with caution',
cv>0.50 ~ 'use with extreme caution',
.default = 'missing or N/A'
))
return(dfs)
}
#' ACS Estimates
#'
#' Generate ACS estimates for multiple tables by tracts, counties, MSAs, or places for multiple years.
#'
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param geography A character string as either 'tract', 'county', 'msa', or 'place'.
#' @param counties A character string or vector of counties. Defaults to PSRC counties.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#' @param FIPS Character string for FIPS codes for specific MSA geographies. Defaults to Seattle & Bremerton MSA c("14740","42660")
#' @param place_FIPS Character string of FIPS codes (with state prefix) for specific Census Places. If NULL, Places within the PSRC Region will be returned.
#' @param acs.type A character string as either 'acs1', 'acs3' or acs5'.
#'
#' @author Craig Helmann
#'
#' @return a tibble of ACS estimates by selected geography for selected table codes. Includes variable names.
#' @examples
#' get_acs_recs(geography = 'county',
#' table.names = c('B03002'),
#' years=c(2017,2019),
#' acs.type = 'acs1')
#'
#' get_acs_recs(geography = 'county',
#' table.names = c("C17002"),
#' counties=c("Kitsap"),
#' years=c(2019),
#' acs.type = 'acs5')
#'
#' @export
get_acs_recs <- function(geography, state="Washington", counties = c('King', 'Kitsap', 'Pierce', 'Snohomish'), table.names, years, FIPS = c("14740","42660"), place_FIPS=NULL, acs.type) {
#this should probably be stored in different place, but I haven't figured out how
if(geography == 'county') {
dfs <- get_acs_county(state, counties, table.names, years, acs.type)
} else if (geography == 'msa'){
dfs <- get_acs_msa(table.names, years, acs.type, FIPS)
} else if(geography == 'place') {
dfs <- get_acs_place(state, table.names, years, acs.type, place_FIPS)
} else if(geography == 'tract') {
dfs <- get_acs_tract(state, counties, table.names, years)
} else if(geography == 'block group') {
dfs <- get_acs_blockgroup(state, counties, table.names, years)
} else if(geography == 'puma') {
dfs <- get_acs_puma(state, table.names, years)
} else if(geography == 'state') {
dfs <- get_acs_state(states=state, table.names, years, acs.type)
}
dfs<-reliability_calcs(dfs)
return(dfs)
}
#' Add shares to Psrccensus ACS object
#'
#' @param df dataframe with Census ACS result
#' @return dataframe with additional share and share_moe fields
#' @author Michael Jensen
#' @rawNamespace import(data.table, except = c(month, year))
#' @export
add_acs_share <- function(df){
label <- x.estimate <- i.estimate <- x.moe <- i.moe <- concept <- share <- share_moe <- NULL
input_type <- class(df)
rs <- setDT(df)
tots <- copy(rs) %>% .[grepl("Total:$",label)]
rs %<>% .[tots, `:=`(share=x.estimate/i.estimate,
share_moe=tidycensus::moe_prop(x.estimate, i.estimate, x.moe, i.moe)),
on=.(GEOID, concept, year)]
if("data.table" %not_in% input_type){rs %<>% setDT()}
return(rs)
}
psrc/psrccensus documentation built on June 1, 2025, 1:06 p.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.
Defines functions add_acs_share get_acs_recs reliability_calcs label_acs_variables acs_varsearch
Documented in acs_varsearch add_acs_share get_acs_recs label_acs_variables reliability_calcs
#' @importFrom magrittr %<>% %>%
NULL
#' Search published ACS variables
#'
#' Identify desired tables by examining prefix of relevant variable codes
#' Includes primary, subject and profile tables
#' @param regex search term
#' @param year optionally restrict search to a specific year
#' @return data.table of filtered variable codes and attributes
#' @author Michael Jensen
#'
#' @rawNamespace import(data.table, except = c(month, year))
#' @importFrom lubridate now month year
#' @export
acs_varsearch <- function(regex, year=NULL){
acs_year <- if(is.null(year)){year(now() - months(18))}else{year}
name <- label <- concept <- NULL # Instantiate tidycensus::pums_variables variable locally (for documentation, not function)
pull_varlist <- function(survey){
x <- tidycensus::load_variables(acs_year, survey) %>% setDT() %>%
.[grepl(regex, name, ignore.case=TRUE)|
grepl(regex, label, ignore.case=TRUE)|
grepl(regex, concept, ignore.case=TRUE)] %>% unique()
return(x)
}
acstypes <- paste0("acs5", c("","/subject","/profile","/cprofile")) %>% c("acsse", recursive=TRUE)
rs <- list()
rs <- lapply(acstypes, pull_varlist) %>% rbindlist(fill=TRUE)
return(rs)
}
#' Label ACS variables
#'
#' Helper function to provide variable labels and concept--i.e. topic--alongside codes
#' @param df dataframe with Census API result
#' @param table.name Census table code
#' @param year the year--or last year--of the ACS survey
#' @param acs.type either acs1 or acs5
#'
#' @return dataframe with labels appended
label_acs_variables <- function(df, table.name, year, acs.type){
name <- label <- concept <- NULL # instantiate variables
survey <- if(table.name %in% acs_tbltypes_lookup$subject){paste0(acs.type,"/subject")
}else if(table.name %in% acs_tbltypes_lookup$profile){paste0(acs.type,"/profile")
}else if(table.name %in% acs_tbltypes_lookup$cprofile){paste0(acs.type,"/cprofile")
}else if(table.name %in% acs_tbltypes_lookup$acsse){"acsse"
}else{acs.type}
labels <- tidycensus::load_variables(year=year, dataset=survey) %>%
dplyr::select(c(name, label, concept)) %>% dplyr::rename(variable=name)
df <- dplyr::left_join(df, labels, by=c("variable"))
return(df)
}
#' ACS Estimates by County
#'
#' Generate ACS estimates for multiple tables by multiple counties
#' This is a helper function for the main function get_acs_recs.
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param counties A character string or vector of counties. Defaults to PSRC counties.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#' @param acs.type A character string as either 'acs1', 'acs3' or acs5'.
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by counties for selected table codes and years with a regional aggregation.
#' Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@importFrom dplyr filter
get_acs_county <- function (state="Washington", counties = c("King","Kitsap","Pierce","Snohomish"), table.names, years, acs.type) {
census.data <- geography <- NAME <- name <- variable <- NULL
estimate <- moe <- sumest <- summoe <- label <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(state=state, geography='county', county=counties, year=year, survey=acs.type, table=table)
# Split County Name for County and State
tbl <- tbl %>% tidyr::separate(col=NAME, into=c("name", "state"),sep=",") %>%
dplyr::mutate(estimate =tidyr::replace_na(estimate,0))
tbl$state <- trimws(tbl$state, "l")
# Create a Total for Region if all 4 PSRC Counties are pulled and then join to county table
if (identical(counties, c("King","Kitsap","Pierce","Snohomish"))){
total <- tbl %>%
dplyr::select(variable, estimate, moe) %>%
dplyr::mutate(estimate =tidyr::replace_na(estimate,0))%>%
dplyr::group_by(variable) %>%
dplyr::summarize(sumest = sum(estimate), summoe = tidycensus::moe_sum(moe, estimate)) %>%
dplyr::rename(estimate=sumest, moe=summoe) %>%
dplyr::mutate(GEOID="REGION", name="Region", state=state)
tbl <- dplyr::bind_rows(list(tbl,total))
}
# Add labels, column for Census Geography, Type and Year of Data
tbl <- tbl %>% label_acs_variables(table, year, acs.type) %>%
dplyr::mutate(census_geography="County", acs_type = acs.type, year=year)
# Median and average calculations are more complicated, may need PUMS, filter out for now:
tbl <- tbl %>% dplyr::filter(name !='Region' |
(!(stringr::str_detect(label, 'Median'))&!(stringr::str_detect(label, 'Average'))))
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by MSA
#'
#' Generate ACS estimates for multiple tables by multiple MSA's
#'
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#' @param acs.type A character string as either 'acs1', 'acs3' or acs5'.
#' @param FIPS Character string for FIPS codes for specific MSA geographies. Defaults to Seattle & Bremerton MSA c("14740","42660")
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by MSA for selected table codes and years. Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@keywords internal
get_acs_msa <- function (table.names, years, acs.type, FIPS = c("14740","42660")) {
census.data <- geography <- NAME <- name <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(geography="metropolitan statistical area/micropolitan statistical area", year=year, survey=acs.type, table=table) %>%
dplyr::filter(GEOID %in% FIPS) %>%
tidyr::separate(col=NAME, into=c("name", "state"),sep=",")
tbl$state <- trimws(tbl$state, "l")
# Add labels, column for Census Geography, Type and Year of Data
tbl %<>% label_acs_variables(table, year, acs.type) %>%
dplyr::mutate(census_geography="MSA", acs_type=acs.type, year=year)
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by Place
#'
#' Generate ACS estimates for multiple tables by multiple places
#'
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#' @param acs.type A character string as either 'acs1', 'acs3' or acs5'.
#' @param place_FIPS Character string of FIPS codes (with state prefix) for specific Census Places. If NULL, Places within the PSRC Region will be returned.
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by Place for selected table codes and years. Includes detailed variable names.
#'
#' @importFrom rlang .data
#' @keywords internal
get_acs_place <- function (state="Washington", table.names, years, acs.type, place_FIPS=NULL) {
census.data <- geography <- NAME <- name <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Determine Places within Region
if(is.null(place_FIPS)){psrc_places <- get_psrc_places(year) %>% sf::st_drop_geometry() %>% unique() %>%
dplyr::pull(GEOID)}
# Download ACS Data
tbl <- tidycensus::get_acs(state=state, geography='place', year=year, survey=acs.type, table=table) %>%
tidyr::separate(col=NAME, into=c("name", "state"),sep=",")
if(!is.null(place_FIPS)){tbl %<>% filter(GEOID %in% place_FIPS)}else{tbl %<>% filter(GEOID %in% psrc_places)}
tbl$state <- trimws(tbl$state, "l")
# Add labels, column for Census Geography, Type and Year of Data
tbl %<>% label_acs_variables(table, year, acs.type) %>%
dplyr::mutate(acs_type = acs.type, year=year) %>%
dplyr::mutate(census_geography = dplyr::case_when(
endsWith(name, "city") ~ "City",
endsWith(name, "CDP") ~ "CDP")) %>%
dplyr::mutate(name = gsub(" city", "", name)) %>%
dplyr::mutate(name = gsub(" CDP", "", name))
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by Census Tract
#'
#' Generate ACS 5yr estimates for multiple tables by Census Tracts in multiple counties
#'
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param counties A character string or vector of counties. Defaults to PSRC counties.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by tracts for selected table codes. Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@keywords internal
get_acs_tract <- function (state="Washington", counties = c("King","Kitsap","Pierce","Snohomish"), table.names, years) {
census.data <- geography <- NAME <- name <- county <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(state=state, county=counties, geography='tract', year=year, survey="acs5", table=table) %>%
tidyr::separate(col=NAME, into=c("name", "county", "state"),sep=",")
tbl$county <- trimws(tbl$county, "l")
tbl$state <- trimws(tbl$state, "l")
county.names <- paste(counties,"County")
# Add label, column for Census Geography, Type and Year of Data
tbl %<>% label_acs_variables(table, year, "acs5") %>%
dplyr::mutate(census_geography="Tract", acs_type = "acs5", year=year) %>%
dplyr::select(-county)
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by Census Block Group
#'
#' Generate ACS 5yr estimates for multiple tables by Census Block Groups in multiple counties
#'
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param counties A character string or vector of counties. Defaults to PSRC counties.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by tracts for selected table codes. Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@keywords internal
get_acs_blockgroup <- function (state="Washington", counties = c("King","Kitsap","Pierce","Snohomish"), table.names, years) {
census.data <- geography <- NAME <- name <- county <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(state=state, county=counties, geography="block group", year=year, survey="acs5", table=table) %>%
tidyr::separate(col=NAME, into=c("name", "tract", "county", "state"),sep=",")
tbl$county <- trimws(tbl$county, "l")
tbl$state <- trimws(tbl$state, "l")
county.names <- paste(counties,"County")
# Add labels, column for Census Geography, Type and Year of Data
tbl %<>% label_acs_variables(table, year, "acs5") %>%
dplyr::mutate(census_geography="block group", acs_type="acs5", year=year) %>%
dplyr::select(-county)
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by Census PUMA
#'
#' Generate ACS 5 year estimates for multiple tables by Census PUMA
#'
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#'
#' @author Christy Lam
#'
#' @return a tibble of acs estimates by PUMA for selected table codes. Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@keywords internal
get_acs_puma <- function (state = "Washington", table.names, years) {
census.data <- geography <- NAME <- name <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(state = state, geography = "public use microdata area", year = year, table = table)
# separate state from PUMA name, clean-up name column
tbl <- tbl %>%
tidyr::separate(col = NAME, into = c("name", "state"), sep = "Washington") %>%
dplyr::mutate(state = 'Washington') %>%
dplyr::mutate(name = stringr::str_replace_all(name, "[\\;|\\,]\\s*$", ""))
# Add labels, column for Census Geography, Type and Year of Data
tbl <- tbl %>%
label_acs_variables(table, year, "acs5") %>%
dplyr::mutate(census_geography = "public use microdata area (PUMA)", acs_type = "acs5", year = year)
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' ACS Estimates by State
#'
#' Generate ACS estimates for multiple tables by state
#' This is a helper function for the main function get_acs_recs.
#' @param states A character string of state names or abbreviations. Defaults to Washington.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#' @param acs.type A character string as either 'acs1', 'acs3' or acs5'.
#'
#' @author Craig Helmann
#'
#' @return a tibble of acs estimates by states for selected table codes.
#' Includes detailed variable names.
#'
#'@importFrom rlang .data
#'@importFrom dplyr filter
get_acs_state <- function (states="Washington", table.names, years, acs.type) {
census.data <- geography <- NAME <- name <- variable <- NULL
estimate <- moe <- sumest <- summoe <- label <- NULL
for (table in table.names) {
yearly.data <- NULL
for (year in years) {
# Download ACS Data
tbl <- tidycensus::get_acs(geography="state", state=states, year=year, survey=acs.type, table=table) %>%
dplyr::mutate(estimate =tidyr::replace_na(estimate,0))
# Format fields for compatibility
tbl <- tbl %>%
tidyr::separate(col = NAME, into = c("name", "state"), sep = "Washington") %>%
dplyr::mutate(state = "Washington", name = "Washington State")
# Add labels, column for Census Geography, Type and Year of Data
tbl <- tbl %>% label_acs_variables(table, year, acs.type) %>%
tidyr::replace_na(list(moe=0)) %>%
dplyr::mutate(census_geography = "state", acs_type = acs.type, year = year)
# Store yearly data into final yearly data for current table - append if a year already exists
ifelse(is.null(yearly.data), yearly.data <- tbl, yearly.data <- dplyr::bind_rows(yearly.data, tbl))
}
# Store table data into final census data - append if a table set already exists
ifelse(is.null(census.data), census.data <- yearly.data, census.data <- dplyr::bind_rows(census.data, yearly.data))
}
return(census.data)
}
#' Add estimate reliability information
#'
#' After gathering data, add reliability information using moe and estimate
#'
#' @param dfs data frames already retrieved from the api, or created by a user
#' @param moe character name of the column that has contains margin of error estimates
#' @param estimate character name of the column that has contains data value estimates
#' @author Suzanne
#' @return the data frames with new columns for se, cv, and reliability
#' @export
reliability_calcs<- function(dfs, moe='moe', estimate='estimate'){
# A coefficient of variation (CV) measures the relative amount of sampling error that is associated with a sample
# estimate. The CV is calculated as the ratio of the SE for an estimate to the estimate itself and is usually
# expressed as a percent.
# Note that since both the ACS and household travel survey are reported using a 90 percent confidence interval,
# where the Margin of Error (MOE) is reported in place of standard error, you can convert it to standard error by
# dividing by 1.645.
# to do: put 1.645 in a better place in the package, it's a magic number
# http://aws-linux/mediawiki/index.php/Understanding_Error_and_Determining_Statistical_Significance
se <- NULL # Bind variables locally (for documentation, not function)
zscore_90<-1.645 # Conversion factor from standard error to 90 percent confidence interval
dfs%<>%dplyr::mutate(se=!!rlang::ensym(moe)/zscore_90)%>%
dplyr::mutate(cv= se/!!rlang::ensym(estimate))%>%
dplyr::mutate(reliability=
dplyr::case_when(
!!rlang::ensym(estimate)==0 ~ 'estimate is 0, cannot compute',
cv<=0.15 ~ 'good',
cv>0.15 & cv<=0.30 ~ 'fair',
cv>0.30 & cv<=0.50 ~ 'use with caution',
cv>0.50 ~ 'use with extreme caution',
.default = 'missing or N/A'
))
return(dfs)
}
#' ACS Estimates
#'
#' Generate ACS estimates for multiple tables by tracts, counties, MSAs, or places for multiple years.
#'
#' @param state A character string state name or abbreviation. Defaults to Washington.
#' @param geography A character string as either 'tract', 'county', 'msa', or 'place'.
#' @param counties A character string or vector of counties. Defaults to PSRC counties.
#' @param table.names A character string or vector of Census table codes.
#' @param years A numeric value or vector of years. An ACS year equal or greater than 2010 to the latest available year.
#' @param FIPS Character string for FIPS codes for specific MSA geographies. Defaults to Seattle & Bremerton MSA c("14740","42660")
#' @param place_FIPS Character string of FIPS codes (with state prefix) for specific Census Places. If NULL, Places within the PSRC Region will be returned.
#' @param acs.type A character string as either 'acs1', 'acs3' or acs5'.
#'
#' @author Craig Helmann
#'
#' @return a tibble of ACS estimates by selected geography for selected table codes. Includes variable names.
#' @examples
#' get_acs_recs(geography = 'county',
#' table.names = c('B03002'),
#' years=c(2017,2019),
#' acs.type = 'acs1')
#'
#' get_acs_recs(geography = 'county',
#' table.names = c("C17002"),
#' counties=c("Kitsap"),
#' years=c(2019),
#' acs.type = 'acs5')
#'
#' @export
get_acs_recs <- function(geography, state="Washington", counties = c('King', 'Kitsap', 'Pierce', 'Snohomish'), table.names, years, FIPS = c("14740","42660"), place_FIPS=NULL, acs.type) {
#this should probably be stored in different place, but I haven't figured out how
if(geography == 'county') {
dfs <- get_acs_county(state, counties, table.names, years, acs.type)
} else if (geography == 'msa'){
dfs <- get_acs_msa(table.names, years, acs.type, FIPS)
} else if(geography == 'place') {
dfs <- get_acs_place(state, table.names, years, acs.type, place_FIPS)
} else if(geography == 'tract') {
dfs <- get_acs_tract(state, counties, table.names, years)
} else if(geography == 'block group') {
dfs <- get_acs_blockgroup(state, counties, table.names, years)
} else if(geography == 'puma') {
dfs <- get_acs_puma(state, table.names, years)
} else if(geography == 'state') {
dfs <- get_acs_state(states=state, table.names, years, acs.type)
}
dfs<-reliability_calcs(dfs)
return(dfs)
}
#' Add shares to Psrccensus ACS object
#'
#' @param df dataframe with Census ACS result
#' @return dataframe with additional share and share_moe fields
#' @author Michael Jensen
#' @rawNamespace import(data.table, except = c(month, year))
#' @export
add_acs_share <- function(df){
label <- x.estimate <- i.estimate <- x.moe <- i.moe <- concept <- share <- share_moe <- NULL
input_type <- class(df)
rs <- setDT(df)
tots <- copy(rs) %>% .[grepl("Total:$",label)]
rs %<>% .[tots, `:=`(share=x.estimate/i.estimate,
share_moe=tidycensus::moe_prop(x.estimate, i.estimate, x.moe, i.moe)),
on=.(GEOID, concept, year)]
if("data.table" %not_in% input_type){rs %<>% setDT()}
return(rs)
}
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.