feat: gains isic_classification Revision 5 (#347)

NEWS.md

@@ -1,6 +1,7 @@
 # r2dii.data (development version)
 
-* Begin deprecation of `cnb_classification` and `isic_classification` (#329).
+* `isic_classification` updated to revision 5 (#329).
+* Begin deprecation of `cnb_classification` (#329).
 * Complete deprecation of `ald_demo` in favor of `abcd_demo` (#328).
 * Complete deprecation of `green_or_brown` in favor of `increasing_or_decreasing` (#307).
 

R/classification_bridge.R

@@ -80,12 +80,16 @@
 #' @inherit nace_classification title
 #' @inherit nace_classification description
 #'
-#' @description
-#' `r lifecycle::badge("deprecated")`
+#' @section Definitions:
+#' `r define('isic_classification')`
 #'
-#' See `?sector_classifications`
+#' @template info_classification-datasets
 #'
-#' @keywords internal
+#' @family datasets for bridging sector classification codes
+#' @seealso [data_dictionary].
+#'
+#' @examples
+#' head(isic_classification)
 "isic_classification"
 
 #' @inherit nace_classification title

data-raw/classification_bridge.R

@@ -7,7 +7,7 @@ nace_classification_raw <- read_bridge(
   file.path("data-raw", "nace_classification.csv")
 )
 
-nace_classification <- convert_superseded_nace_code(
+nace_classification <- prepend_letter_nace_code(
   nace_classification_raw,
   col_from = "original_code",
   col_to = "code"
@@ -34,3 +34,53 @@ psic_classification <- read_bridge(
   file.path("data-raw", "psic_classification.csv")
 )
 usethis::use_data(psic_classification, overwrite = TRUE)
+
+isic_classification_raw <- read_csv_(
+  file.path("data-raw", "isic_classification_rev_5.csv")
+)
+
+isic_classification <- prepend_letter_isic_code(
+  isic_classification_raw,
+  col_from = "ISIC Rev 5 Code",
+  col_to = "code"
+)
+
+isic_classification <- dplyr::mutate(
+  isic_classification,
+  sector = dplyr::case_when(
+    grepl("^B05", code) ~ "coal",
+    grepl("^B06", code) ~ "oil and gas",
+    grepl("^B091", code) ~ "oil and gas", # borderline
+    grepl("^B099", code) ~ "coal", # borderline
+    grepl("^C2394", code) ~ "cement",
+    grepl("^C2395", code) ~ "cement", # borderline
+    grepl("^C241", code) ~ "steel",
+    grepl("^C2431", code) ~"steel", # borderline
+    grepl("^C291", code) ~ "automotive", # borderline
+    grepl("^C292", code) ~ "automotive", # borderline
+    grepl("^C293", code) ~ "automotive", # borderline
+    grepl("^D351", code) ~ "power", # some of these are borderline
+    grepl("^H50", code) ~ "shipping",
+    grepl("^H51", code) ~ "aviation",
+    TRUE ~ "not in scope"
+    ),
+  borderline = dplyr::case_when(
+    grepl("^B091", code) ~ TRUE,
+    grepl("^B099", code) ~ TRUE,
+    grepl("^C2395", code) ~ TRUE,
+    grepl("^C2431", code) ~ TRUE,
+    grepl("^C291", code) ~ TRUE,
+    grepl("^C292", code) ~ TRUE,
+    grepl("^C293", code) ~ TRUE,
+    code == "D351" ~ TRUE,
+    grepl("^D3513", code) ~ TRUE,
+    TRUE ~ FALSE
+  ),
+)
+
+isic_classification <- dplyr::mutate(
+  isic_classification,
+  revision = "5"
+)
+
+usethis::use_data(isic_classification, overwrite = TRUE)

data-raw/data_dictionary/isic_classification.csv

@@ -1,5 +1,7 @@
 dataset,column,typeof,definition
-isic_classification,code,character,Original ISIC code
-isic_classification,code_level,double,Level of granularity of ISIC code
-isic_classification,sector,character,Associated 2dii sector
-isic_classification,borderline,logical,Flag indicating if 2dii sector and classification code are a borderline match. The value TRUE indicates that the match is uncertain between the 2dii sector and the classification. The value FALSE indicates that the match is certainly perfect or the classification is certainly out of 2dii's scope.
+isic_classification,ISIC Rev 5 Code,character,Original ISIC Rev 5 code
+isic_classification,ISIC Rev 5 Title,character,Original ISIC Rev 5 title
+isic_classification,code,character,ISIC Rev 5 code with top-level letter prepended
+isic_classification,sector,character,Associated PACTA sector
+isic_classification,borderline,logical,Flag indicating if PACTA sector and classification code are a borderline match. The value TRUE indicates that the match is uncertain between the PACTA sector and the classification. The value FALSE indicates that the match is certainly perfect or the classification is certainly out of PACTA's scope.
+isic_classification,revision,character,Column identifying to which ISIC revision the code belongs.