Wagebill analytics

.github/workflows/R-CMD-check.yaml

@@ -2,6 +2,7 @@
 # Need help debugging build failures? Start at https://github.com/r-lib/actions#where-to-find-help
 on:
   push:
+    branches: [main, master]
   pull_request:
 
 name: R-CMD-check.yaml

.gitignore

@@ -4,3 +4,4 @@
 .Ruserdata
 *.rds
 docs
+spielplatz/

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: govhr
 Type: Package
 Title: Clean and analyze data that informs human resource management
-Version: 0.3.5
+Version: 0.4.0
 Description: This package is a toolkit developed by the World Bank. It cleans, harmonizes, and analyzes human resource data with a standardized methodology.
 License: MIT + file LICENSE
 Encoding: UTF-8
@@ -35,9 +35,11 @@ Imports:
     tidygraph,
     ggraph,
     ggiraph,
-    igraph
-Suggests: 
+    igraph,
+    fixest,
     forcats,
+    ggstats
+Suggests: 
     readr,
     testthat (>= 3.0.0),
     reactable
@@ -50,4 +52,5 @@ Authors@R: c(
     person("Galileu", "Kim", , "galileukim@worldbank.org", role = c("aut", "cre")),
     person("Ifeanyi", "Edochie", , "iedochie@worldbank.org", role = c("aut"))
   )
+Roxygen: list(markdown = TRUE)
 Config/roxygen2/version: 8.1.0

NAMESPACE

@@ -19,15 +19,18 @@ export(compute_fastsummary)
 export(compute_global_consistency)
 export(compute_global_coverage)
 export(compute_growth)
+export(compute_growth_decomposition)
 export(compute_growth_summary)
 export(compute_movement_cost)
 export(compute_pension_ratio)
 export(compute_percentile)
 export(compute_quantile)
 export(compute_record_consistency)
+export(compute_service_table)
 export(compute_time_trend)
 export(compute_trend_summary)
 export(compute_value_consistency)
+export(compute_wage_decomposition)
 export(compute_wagebill)
 export(compute_workforce_movement)
 export(convert_constant_ppp)
@@ -39,11 +42,11 @@ export(dedup_values)
 export(define_fns)
 export(deflate_to_real)
 export(detect_career_transition)
-export(detect_career_transitions)
 export(detect_inconsistent_cols)
 export(detect_personnel_event)
 export(detect_reallocation)
 export(detect_retirement)
+export(estimate_decrement_rates)
 export(estimate_exit_rates)
 export(fastcount)
 export(fastprop)
@@ -64,6 +67,7 @@ export(ggplot_point_line)
 export(ggplot_segment)
 export(guess_date_frequency)
 export(harmonize_columns)
+export(model_wage)
 export(pivot_data360)
 export(plot_bar_growth)
 export(plot_bar_total)
@@ -75,6 +79,8 @@ export(plot_coverage_heatmap)
 export(plot_coverage_trend)
 export(plot_decile)
 export(plot_histogram)
+export(plot_model_wage)
+export(plot_model_wage_fes)
 export(plot_movement)
 export(plot_movement_cost)
 export(plot_segment)
@@ -86,14 +92,16 @@ export(remove_duplicate_personnel)
 export(rescale_baseline)
 export(sample_group)
 export(scale_plot_height)
+export(smooth_decrement_rates)
 export(validate_data)
 import(dplyr)
 import(ggplot2)
-import(glue)
 import(httr)
 importFrom(broom,tidy)
 importFrom(collapse,fquantile)
 importFrom(data.table,
+  ":=",
+  .N,
   CJ,
   as.data.table,
   copy,
@@ -120,7 +128,6 @@ importFrom(dplyr,
   any_of,
   arrange,
   case_when,
-  collect,
   count,
   desc,
   distinct,
@@ -150,6 +157,8 @@ importFrom(dplyr,
   union
 )
 importFrom(dtplyr,lazy_dt)
+importFrom(fixest,fixef)
+importFrom(forcats,fct_reorder)
 importFrom(ggiraph,
   geom_point_interactive,
   girafe,
@@ -158,8 +167,8 @@ importFrom(ggiraph,
 )
 importFrom(ggplot2,
   aes,
-  arrow,
   coord_cartesian,
+  element_text,
   expansion,
   facet_wrap,
   geom_col,
@@ -172,9 +181,9 @@ importFrom(ggplot2,
   geom_vline,
   ggplot,
   guide_axis,
+  label_wrap_gen,
   labs,
   margin,
-  position_jitter,
   scale_color_manual,
   scale_size_identity,
   scale_x_continuous,
@@ -195,10 +204,10 @@ importFrom(ggraph,
   scale_edge_width_continuous
 )
 importFrom(ggrepel,geom_text_repel)
-importFrom(ggthemes,scale_color_few)
+importFrom(ggstats,ggcoef_model)
+importFrom(ggthemes,scale_colour_few)
 importFrom(glue,glue)
 importFrom(grDevices,colorRampPalette)
-importFrom(grid,unit)
 importFrom(httr,
   GET,
   content,
@@ -262,16 +271,9 @@ importFrom(tibble,
 )
 importFrom(tidygraph,as_tbl_graph)
 importFrom(tidyr,
+  complete,
   nest,
-  pivot_longer,
+  nesting,
   pivot_wider
 )
 importFrom(tidyselect,eval_select)
-importFrom(validate,
-  confront,
-  description,
-  label,
-  summary,
-  validator,
-  values
-)

NEWS.md

@@ -1,3 +1,11 @@
+# govhr 0.4.0
+This release:
+- Introduces novel wage bill modelling and demographic analysis.
+- Renames the function arguments, ensuring consistency.
+- Warns on deprecated argument names, to be removed in the next relase.
+- Fixes documentation and their rendering.
+- Fixes invalid and unused imports.
+
 # govhr 0.3.5
 This release:
 - Ports data transformation and plotting functions from govhrapp to govhr.

R/qcheck_consistency.R → R/consistency.R

@@ -1,7 +1,7 @@
 #' Compute the proportion of consistent records and values in a data frame.
 #'
 #' @param data A data frame.
-#' @param id_col A character string specifying the name of the column that uniquely identifies records.
+#' @param id_col A string specifying the name of the column that uniquely identifies records.
 #' @param value_cols A character vector specifying the name(s) of columns whose values
 #'   are to be checked for consistency. Value consistency is computed separately for
 #'   each column and averaged across columns before being combined with record consistency.
@@ -10,7 +10,7 @@
 #' @import dplyr
 #' @importFrom purrr map_dbl
 #'
-#' @return A numeric value representing the proportion of consistent records and values in the data frame.
+#' @returns A numeric value representing the proportion of consistent records and values in the data frame.
 #' @details Consistency is defined as the proportion of records and values that are consistent
 #'   across the dataset. A record is considered consistent if it has a unique identifier and all
 #'   its associated values are consistent. A value is considered consistent if it does not
@@ -57,18 +57,18 @@ compute_global_consistency <- function(data, id_col, value_cols, digits = 2) {
     round(digits)
 }
 
-#' Compute the proportion of consistent records in a data frame.
+#' Compute the proportion of consistent records in a data frame
 #'
 #' @param data A data frame.
-#' @param id_col A character string specifying the name of the column that uniquely identifies records (e.g., "personnel_id" or "contract_id").
+#' @param id_col A string specifying the name of the column that uniquely identifies records (e.g., "personnel_id" or "contract_id").
 #' @param group_cols A character vector specifying the names of the columns to group by. Default is NULL, which means no grouping.
 #' @param digits An integer specifying the number of decimal places to round the result to. Default is 2.
 #'
 #' @import dplyr
 #' @importFrom data.table as.data.table fifelse
 #' @importFrom tibble as_tibble
 #'
-#' @return A data frame with the proportion of consistent records in the data frame, optionally by group.
+#' @returns A data frame with the proportion of consistent records in the data frame, optionally by group.
 #'
 #' @details A record is considered consistent if it has a unique identifier and all its associated values are consistent.
 #' The function computes the proportion of consistent records in the data frame, optionally grouped by specified columns.
@@ -130,18 +130,18 @@ compute_record_consistency <- function(
   tibble::as_tibble(result)
 }
 
-#' Compute the proportion of consistent values in a data frame.
+#' Compute the proportion of consistent values in a data frame
 #'
 #' @param data A data frame.
-#' @param id_col A character string specifying the name of the column that uniquely identifies records.
-#' @param value_col A character string specifying the name of the column whose values are to be checked for consistency.
+#' @param id_col A string specifying the name of the column that uniquely identifies records.
+#' @param value_col A string specifying the name of the column whose values are to be checked for consistency.
 #' @param group_cols A character vector specifying the names of the columns to group by. Default is no grouping.
 #' @param digits An integer specifying the number of decimal places to round the result to. Default is 2.
 #'
 #' @importFrom data.table as.data.table
 #' @importFrom tibble as_tibble
 #'
-#' @return A data frame with the proportion of consistent values in the data frame, optionally by group.
+#' @returns A data frame with the proportion of consistent values in the data frame, optionally by group.
 #'
 #' @details Consistency is broadly defined as the proportion of records and values that are consistent
 #'   across the dataset. A value is considered consistent if it does not differ from

R/qcheck_coverage.R → R/coverage.R

@@ -1,49 +1,52 @@
 #' Compute coverage of non-missing values in a dataset.
 #'
-#' @param .data A data frame.
-#' @param group A character string specifying the column name to group by. If NULL, coverage is computed for the entire data set.
+#' @param data A data frame.
+#' @param group_cols A string specifying the column name to group by. If NULL, coverage is computed for the entire data set.
 #' @param include_ref_date A logical value indicating whether to include the `ref_date` column in the grouping.
 #' @param aggregate A logical value indicating whether to aggregate coverage values by the `group`.
+#' @param group Deprecated. Use `group_cols` instead.
 #'
 #' @importFrom data.table as.data.table
 #' @importFrom tibble as_tibble
 #'
-#' @return A data frame with coverage values for each column, optionally grouped by the specified `group`.
+#' @returns A data frame with coverage values for each column, optionally grouped by the specified `group`.
 #' 
 #' @export
 compute_coverage <- function(
-  .data,
-  group = NULL,
+  data,
+  group_cols = NULL,
   include_ref_date = FALSE,
-  aggregate = FALSE
+  aggregate = FALSE,
+  group = NULL
 ) {
-  dt <- data.table::as.data.table(.data)
+  group_cols <- resolve_renamed_arg(group_cols, group, "group", "group_cols")
+  dt <- data.table::as.data.table(data)
   data_cols <- colnames(dt)
 
   if (include_ref_date) {
-    group <- unique(c("ref_date", group))
+    group_cols <- unique(c("ref_date", group_cols))
   }
 
-  summary_cols <- setdiff(data_cols, group)
+  summary_cols <- setdiff(data_cols, group_cols)
 
   # wide: one coverage value per summary column, one row per group
-  if (is.null(group)) {
+  if (is.null(group_cols)) {
     coverage_wide <- dt[,
       lapply(.SD, \(col) (sum(!is.na(col)) / length(col)) * 100),
       .SDcols = summary_cols
     ]
   } else {
     coverage_wide <- dt[,
       lapply(.SD, \(col) (sum(!is.na(col)) / length(col)) * 100),
-      by = c(group),
+      by = c(group_cols),
       .SDcols = summary_cols
     ]
   }
 
   # long: pivot summary_cols into variable/coverage pairs
   coverage_data <- data.table::melt(
     coverage_wide,
-    id.vars = group,
+    id.vars = group_cols,
     measure.vars = summary_cols,
     variable.name = "variable",
     value.name = "coverage",
@@ -53,7 +56,7 @@ compute_coverage <- function(
   if (aggregate) {
     coverage_data <- coverage_data[,
       .(coverage = mean(coverage, na.rm = TRUE)),
-      by = c(group)
+      by = c(group_cols)
     ]
   }
 
@@ -65,7 +68,7 @@ compute_coverage <- function(
 #' @param data A data frame.
 #' @param digits An integer specifying the number of decimal places to round the result to. Default is 2.
 #'
-#' @return A numeric value representing the proportion of missing values in the data frame.
+#' @returns A numeric value representing the proportion of missing values in the data frame.
 #' 
 #' @export
 compute_global_coverage <- function(data, digits = 2) {

R/data.R

@@ -367,6 +367,7 @@
 #'   \item{race}{Worker's race or broad ethnic classification, where available.}
 #'   \item{tribe}{Worker's ethnic or tribal affiliation, where available.}
 #'   \item{first_employment_date}{Date the worker first entered the public service.}
+#'   \item{retirement_date}{The date of retirement for those whose \code{employment_status} is \code{"pensioner"}}
 #' }
 #'
 #' @details

R/data360api.R

@@ -1,6 +1,6 @@
 #' This function retrieves data from the Data360 API.
 #'
-#' @return A tibble containing data, including country codes, country names,
+#' @returns A tibble containing data, including country codes, country names,
 #'         years, and values.
 #' @examples
 #' \dontrun{
@@ -16,7 +16,7 @@
 #' @import dplyr
 #' @importFrom janitor clean_names
 #' @importFrom jsonlite fromJSON
-#' @importFrom tibble as_tibble tibble
+#' @importFrom tibble tibble
 #' @export
 get_data360_api <- function(dataset_id, indicator_id, pivot = TRUE) {
   base_url <- "https://data360api.worldbank.org/data360/data"
@@ -89,7 +89,7 @@ get_data360_api <- function(dataset_id, indicator_id, pivot = TRUE) {
 #'   - `INDICATOR`: Indicator code or name
 #'   - `OBS_VALUE`: Observation value for the indicator
 #'
-#' @return A tibble in wide format with columns:
+#' @returns A tibble in wide format with columns:
 #'   - `country_code`: The country or region code (from `REF_AREA`)
 #'   - `year`: The year or time period (from `TIME_PERIOD`)
 #'   - One column per unique `INDICATOR`, containing corresponding values from `OBS_VALUE`
@@ -143,10 +143,10 @@ pivot_data360 <- function(data) {
 #' URL, retrieves the metadata in JSON format, and parses it into an R list or
 #' data frame.
 #'
-#' @param dataset_id A character string or numeric identifier specifying the
+#' @param dataset_id A string or numeric identifier specifying the
 #'   dataset for which metadata should be retrieved.
 #'
-#' @return A list (or data frame) containing the metadata associated with the
+#' @returns A list (or data frame) containing the metadata associated with the
 #'   requested dataset, as returned by the Data360 API.
 #'
 #' @examples