In unicef-drp/CME.assistant: Reusable Assisting Functions for Data Analysis

knitr::opts_chunk$set(
  message = FALSE, warning = FALSE, 
  collapse = TRUE,
  comment = "#>"
)

The goal of CME.assistant is to store frequently used functions within the CM team.

options(rmarkdown.html_vignette.check_title = FALSE)

Functions for reading data

For now the most useful ones I used in many scripts are:

read.results.csv: read any "results.csv"
read.country.summary: read any "Rates & Deaths_Country Summary.csv"
read.region.summary: read any "Rates & Deaths_...Region.csv"

Examples:

# reproduce these code requires access to our internal Dropbox folders

# Always a good idea to update the library to make sure you have the latest version
# devtools::install_github("unicef-drp/CME.assistant")
# library("CME.assistant")
USERPROFILE <- CME.assistant::load_os_leading_dir() # leading dir to Dropbox

# Dropbox directories to all results.csv
dir_CC_code <- file.path(USERPROFILE, "Dropbox/UNICEF Work/Country consultation/Code_for_CC")
source(file.path(dir_CC_code, "R/Dropbox_results_directories_2021.R"))
dt_results <- rbindlist(lapply(results_dir_list_final_2021, CME.assistant::read.results.csv))
# which the same as:
dt_results <- do.call(rbind, lapply(results_dir_list_final_2021, CME.assistant::read.results.csv))
dt_results[!is.na(value), table(Sex, Shortind)]


# Dropbox directories to all final aggregates
dir_report_code <- file.path(USERPROFILE, "Dropbox/UNICEF Work/IGME report etc/2021/Code_for_report/")
source(file.path(dir_report_code, "Dropbox_aggresults_directories_2021.R"))
dir_country_summary <- c(
  file.path(dir_aggu5,      "Rates & Deaths_Country Summary.csv"),
  file.path(dir_aggu5_f,    "Rates & Deaths(ADJUSTED)_female_Country Summary.csv"),
  file.path(dir_aggu5_m,    "Rates & Deaths(ADJUSTED)_male_Country Summary.csv"),
  file.path(dir_agg10q5,    "Rates & Deaths_Country Summary.csv"),
  file.path(dir_agg10q5_f,  "Rates & Deaths(ADJUSTED)_Country Summary.csv"),
  file.path(dir_agg10q5_m,  "Rates & Deaths(ADJUSTED)_Country Summary.csv"),
  file.path(dir_agg10q15,   "Rates & Deaths_Country Summary.csv"),
  file.path(dir_agg10q15_f, "Rates & Deaths(ADJUSTED)_Country Summary.csv"),
  file.path(dir_agg10q15_m, "Rates & Deaths(ADJUSTED)_Country Summary.csv")
)
dt_estimates <- rbindlist(lapply(dir_country_summary, CME.assistant::read.country.summary))
dt_estimates[, table(Shortind, Sex)]

Similar for regional summary:

region_group_filename <- "SDGSimpleRegion"
dir_region_summary <- c(
  file.path(dir_aggu5,   paste0("Rates & Deaths_", region_group_filename, ".csv")),
  file.path(dir_aggu5_f, paste0("Rates & Deaths(ADJUSTED)_female_", region_group_filename, ".csv")),
  file.path(dir_aggu5_m, paste0("Rates & Deaths(ADJUSTED)_male_", region_group_filename, ".csv")),
  file.path(dir_agg10q5,   paste0("Rates & Deaths_", region_group_filename, ".csv")),
  file.path(dir_agg10q15,   paste0("Rates & Deaths_", region_group_filename, ".csv"))
)
dt_region <- rbindlist(lapply(dir_region_summary, CME.assistant::read.region.summary))

get.CME.UI.data: A more advanced version of read.country.summary that allows more tuning: can output wide format by wide-quantile, wide-year, wide-indicator or one column for rate and one for death.
The default long-format output contains columns c("ISO3Code", "OfficialName", "Shortind", "Year", "Quantile", "Sex", "value")
The function will output all available indicators in the dataset
Could subset by ISO (c_iso) or year (year_range)

Examples:

dir_cs_u5 <- file.path(dir_aggu5, "Rates & Deaths_Country Summary.csv")
dt_1 <- get.CME.UI.data(dir_file = dir_cs_u5)
dt_1[Year == 2020][1:3,]
dt_1 <- get.CME.UI.data(dir_file = dir_cs_u5, 
                        idvars = c("ISO3Code", "CountryName", "OfficialName"), format = "wide_q")
dt_1[Year == 2020][1,]
dt_1 <- get.CME.UI.data(dir_file = dir_cs_u5, format = "wide_q")
dt_1[Year == 2020][1,]
dt_1 <- get.CME.UI.data(dir_file = dir_cs_u5, format = "wide_ind", round_digit = 1)
dt_1[Year == 2020][1:3,]
dt_1 <- get.CME.UI.data(dir_file = dir_cs_u5, format = "wide_get", round_digit = 1)
dt_1[Year == 2020][1:3,]
dt_wy <- get.CME.UI.data(dir_file = dir_cs_u5, format = "wide_year", year_range = c(2000, 2010, 2020))
dt_wy[1:3,]

Functions for calculating

Using the wide-year output above, we can directly calculate ARR or percentage decline (PD)

dt_wy <- calculate.arr(dt_wy, 2000, 2010) # ARR
dt_wy <- calculate.arr(dt_wy, 2010, 2020) # ARR
dt_wy <- calculate.pd(dt_wy, 2000, 2020) # percentage decline
dt_wy[Quantile == "Median" & Shortind == "NMR", ][1:3,]

Locate the latest databases

Automatically point to the latest databases judging by the file names

Sometimes it is good to write down the directory of specific version of database. But these could be helpful when we just need to load the latest databases.

dir_IGME_input <- get.IGMEinput.dir(2022)
dir_U5MR   <- get.dir_U5MR(dir_IGME_input)
dir_IMR    <- get.dir_IMR(dir_IGME_input)
dir_NMR    <- get.dir_NMR(y5 = TRUE) # either 5-year or not
dir_NMR    <- get.dir_NMR(y5 = FALSE) # either 5-year or not
dir_gender <- get.dir_gender(plotting = TRUE) # either dataset for plotting or modeling
dir_gender <- get.dir_gender(plotting = FALSE) # either dataset for plotting or modeling