Merge pull request #103 from peteowen1/anzsic-2006-fix

update anzsic2006 to download from abs source.

DESCRIPTION

@@ -11,7 +11,8 @@ Authors@R: c(person("Will", "Mackey", email = "[email protected]", role = c("au
         person("Benjamin", "Wee", role = c("aut")),
         person("Carlos", "Yanez", role = "ctb"),
         person("Bas", "Latcham", role = "ctb"),
-        person("Rex", "Parsons", role = "ctb", comment = c(ORCID = "0000-0002-6053-8174"))
+        person("Rex", "Parsons", role = "ctb", comment = c(ORCID = "0000-0002-6053-8174")),
+        person("Pete", "Owen", role = "ctb")
         )
 Maintainer: Will Mackey <[email protected]>
 License: GPL-3

NEWS.md

@@ -1,6 +1,7 @@
 # strayr (development version)
 * `create read_correspondence_tbl()` reads correspondence tables from
 `absmapsdata` similarly to `read_absmap()`
+* updated `anzsco2006` to include leading zeros in codes (see ). This is a backwards incompatible change that may cause issues (not enough for a major version progression)
 
 # strayr 0.2.2
 * `anzsco2022` updated to reflect changes made by the ABS

R/data_descriptions.R

@@ -21,10 +21,10 @@
 "anzsco2009"
 
 
-#' ANZSCO 2019
+#' ANZSCO 2013
 #'
 #' Wide table containing all levels of the Australian and New Zealand Standard
-#' Classification of Occupations (ANZSCO), Version 1.3, 2019
+#' Classification of Occupations (ANZSCO), Version 1.2, 2013
 #'
 #' @format A \code{tibble} with 11 variables:
 #' \describe{
@@ -41,14 +41,13 @@
 #' \item{\code{skill_level}}{Skill level required for occupation, determined by the ABS (1 is highest, 5 is lowest).
 #' See \url{https://www.abs.gov.au/ausstats/[email protected]/Previousproducts/C4BECE1704987586CA257089001A9181 } for details.}
 #' }
-"anzsco2019"
-
+"anzsco2013"
 
 
-#' ANZSCO 2013
+#' ANZSCO 2019
 #'
 #' Wide table containing all levels of the Australian and New Zealand Standard
-#' Classification of Occupations (ANZSCO), Version 1.2, 2013
+#' Classification of Occupations (ANZSCO), Version 1.3, 2019
 #'
 #' @format A \code{tibble} with 11 variables:
 #' \describe{
@@ -65,7 +64,8 @@
 #' \item{\code{skill_level}}{Skill level required for occupation, determined by the ABS (1 is highest, 5 is lowest).
 #' See \url{https://www.abs.gov.au/ausstats/[email protected]/Previousproducts/C4BECE1704987586CA257089001A9181 } for details.}
 #' }
-"anzsco2013"
+"anzsco2019"
+
 
 #' ANZSCO 2021
 #'
@@ -89,6 +89,7 @@
 #' }
 "anzsco2021"
 
+
 #' ANZSCO 2022
 #'
 #' Wide table containing all levels of the Australian and New Zealand Standard
@@ -113,28 +114,26 @@
 "anzsco2022"
 
 
-#' ANZSIC
+#' ANZSIC 2006
 #'
 #' Wide table containing all levels of the Australian and New Zealand Standard
-#' Industrial Classification (ANZSIC), 2006 (Revision 1.0). Cat. 1292.0.
+#' Industrial Classification (ANZSIC), 2006 (Revision 2.0). Cat. 1292.0.
 #'
 #' @format A \code{tibble} with 8 variables:
 #' \describe{
 #' \item{\code{anzsic_division_code}}{ANZSIC division codes character, e.g. "A", "B"}
 #' \item{\code{anzsic_division}}{ANZSIC division title, e.g. "Agriculture, Forestry and Fishing"}
-#' \item{\code{anzsic_subdivision_code}}{ANZSIC subdivision codes integer, e.g. 1, 2}
+#' \item{\code{anzsic_subdivision_code}}{ANZSIC subdivision codes 2-digit character, e.g. 01, 02}
 #' \item{\code{anzsic_subdivision}}{ANZSIC subdivision title, e.g. "Agriculture"}
-#' \item{\code{anzsic_group_code}}{ANZSIC group codes integer, e.g. 11, 12}
+#' \item{\code{anzsic_group_code}}{ANZSIC group codes 3-digit character, e.g. 011, 012}
 #' \item{\code{anzsic_group}}{ANZSIC group title, e.g. "Mushroom and Vegetable Growing"}
-#' \item{\code{anzsic_class_code}}{ANZSIC class codes integer, e.g. 111, 112}
+#' \item{\code{anzsic_class_code}}{ANZSIC class codes 4-digit character, e.g. 0111, 0112}
 #' \item{\code{anzsic_class}}{ANZSIC class title, e.g. "Vegetable Growing (Under Cover)"}
 #' }
+#' @source \url{https://www.abs.gov.au/statistics/classifications/australian-and-new-zealand-standard-industrial-classification-anzsic/2006-revision-2-0/numbering-system-and-titles/division-subdivision-group-and-class-codes-and-titles}
 "anzsic2006"
 
 
-
-
-
 #' ASCED Field of Education
 #'
 #' Wide table containing all levels of fields of education in the Australian

README.Rmd

@@ -69,7 +69,7 @@ Current structures stored in `strayr` are:
     - `anzsco2013`: occupation levels of ANZSCO, [2013, Version 1.2](https://www.abs.gov.au/AUSSTATS/[email protected]/allprimarymainfeatures/4AF138F6DB4FFD4BCA2571E200096BAD?opendocument).
     - `anzsco2009`: occupation levels ANZSCO, [First Edition, Revision 1, 2009](https://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/1220.0First%20Edition,%20Revision%201?OpenDocument).
 - Australian and New Zealand Standard Industrial Classification (**ANZSIC**), Cat. 1292.0:
-    - `anzsic2006`: industry levels of ANZSIC, [2006 (Revision 1.0)](https://www.abs.gov.au/ausstats/[email protected]/0/20C5B5A4F46DF95BCA25711F00146D75?opendocument).
+    - `anzsic2006`: industry levels of ANZSIC, [2006 (Revision 2.0)](https://www.abs.gov.au/statistics/classifications/australian-and-new-zealand-standard-industrial-classification-anzsic/2006-revision-2-0).
 - Australian Standard Classification of Education (**ASCED**), Cat. 1272.0:
     - `asced_foe2001`: field of education levels of ASCED, [2001](https://www.abs.gov.au/ausstats/[email protected]/mf/1272.0).
     - `asced_qual2001`: qualification levels of ASCED, [2001](https://www.abs.gov.au/ausstats/[email protected]/mf/1272.0).

README.md

@@ -60,7 +60,7 @@ Current structures stored in `strayr` are:
 - Australian and New Zealand Standard Industrial Classification
   (**ANZSIC**), Cat. 1292.0:
   - `anzsic2006`: industry levels of ANZSIC, [2006 (Revision
-    1.0)](https://www.abs.gov.au/ausstats/[email protected]/0/20C5B5A4F46DF95BCA25711F00146D75?opendocument).
+    2.0)](https://www.abs.gov.au/statistics/classifications/australian-and-new-zealand-standard-industrial-classification-anzsic/2006-revision-2-0).
 - Australian Standard Classification of Education (**ASCED**), Cat.
   1272.0:
   - `asced_foe2001`: field of education levels of ASCED,

data-raw/create_anzsic2006.R

@@ -2,47 +2,78 @@
 # Reading and cleaning ANZSIC correspondence
 
 library(tidyverse)
-library(glue)
+library(rvest)
 
 # include factor variants or nah?
 include_factor_variants <- FALSE
 
-# ty asiripanich
-anzsic_url <- "https://raw.githubusercontent.com/asiripanich/anzsic/master/anzsic_2006.csv"
-
-# Read
-anzsic_raw <- read_csv(anzsic_url) %>%
-    rename_all(~ glue("anzsic_{.}")) %>%
-    mutate_if(is.double, as.integer) %>%
-    as_tibble()
-
-# Add layers of nfd
-class_nfd <- anzsic_raw %>%
-  distinct(anzsic_division_title, anzsic_division_code,
-           anzsic_subdivision_title, anzsic_subdivision_code,
-           anzsic_group_title, anzsic_group_code) %>%
-  mutate(anzsic_class_code = anzsic_group_code * 10,
-         anzsic_class_title = glue("{anzsic_group_title}, nfd"))
-
-group_nfd <- anzsic_raw %>%
-  distinct(anzsic_division_title, anzsic_division_code,
-           anzsic_subdivision_title, anzsic_subdivision_code) %>%
-  mutate(anzsic_group_title = glue("{anzsic_subdivision_title}, nfd"),
-         anzsic_group_code = anzsic_subdivision_code * 10,
-         anzsic_class_title = anzsic_group_title,
-         anzsic_class_code = anzsic_group_code * 10)
-
-subdivision_nfd <- anzsic_raw %>%
-  group_by(anzsic_division_code, anzsic_division_title) %>%
-  summarise(anzsic_subdivision_code = min(anzsic_subdivision_code)) %>%
-  mutate(anzsic_subdivision_title = glue("{anzsic_division_title}, nfd"),
-         anzsic_group_title = anzsic_subdivision_title,
-         anzsic_group_code = anzsic_subdivision_code * 10,
-         anzsic_class_title = anzsic_group_title,
-         anzsic_class_code = anzsic_group_code * 10)
+# fetch from abs website
+url <- "https://www.abs.gov.au/statistics/classifications/australian-and-new-zealand-standard-industrial-classification-anzsic/2006-revision-2-0/numbering-system-and-titles/division-subdivision-group-and-class-codes-and-titles"
+
+df <- url %>%
+  rvest::read_html() %>%
+  rvest::html_table()
+
+# bind tables together
+anzsic_2006_temp <-
+  purrr::list_rbind(df)
+
+# fix columns names and bind together
+colnames(anzsic_2006_temp) <- c("anzsic_division_code", "anzsic_subdivision_code", "anzsic_group_code", "anzsic_class_code", "title")
+
+first_row <-
+  as.data.frame(t(colnames(df[[1]])))
+
+colnames(first_row) <- c("anzsic_division_code", "anzsic_subdivision_code", "anzsic_group_code", "anzsic_class_code", "title")
+
+anzsic_2006_total <-
+  dplyr::bind_rows(first_row, anzsic_2006_temp)
+
+# replace blanks with NAs
+anzsic_2006_total[anzsic_2006_total == ""] <- NA
+
+# fill NAs down from above
+anzsic_2006_fill <-
+  anzsic_2006_total %>%
+  tidyr::fill(colnames(anzsic_2006_total), .direction = c("down"))
+
+# get each grouping type individually
+anzsic_2006_class <-
+  anzsic_2006_total %>%
+  dplyr::filter(stringr::str_detect(anzsic_class_code, "^[:digit:]+$")) %>%
+  dplyr::select(anzsic_class_code, anzsic_class_title = title)
+
+anzsic_2006_group <-
+  anzsic_2006_total %>%
+  dplyr::filter(stringr::str_detect(anzsic_group_code, "^[:digit:]+$")) %>%
+  dplyr::select(anzsic_group_code, anzsic_group_title = title)
+
+anzsic_2006_subdivision <-
+  anzsic_2006_total %>%
+  dplyr::filter(stringr::str_detect(anzsic_subdivision_code, "^[:digit:]+$")) %>%
+  dplyr::select(anzsic_subdivision_code, anzsic_subdivision_title = title)
+
+anzsic_2006_division <-
+  anzsic_2006_total %>%
+  dplyr::filter(stringr::str_detect(anzsic_division_code, "^[:alpha:]+$")) %>%
+  dplyr::select(anzsic_division_code, anzsic_division_title = title)
+
+# combine grouping types into final table
+anzsic_2006_final <-
+  anzsic_2006_fill %>%
+  dplyr::left_join(anzsic_2006_division) %>%
+  dplyr::left_join(anzsic_2006_subdivision) %>%
+  dplyr::left_join(anzsic_2006_group) %>%
+  dplyr::left_join(anzsic_2006_class) %>%
+  dplyr::filter(!is.na(anzsic_class_title)) %>%
+  dplyr::select(
+    anzsic_division_code, anzsic_division_title, anzsic_subdivision_code, anzsic_subdivision_title,
+    anzsic_group_code, anzsic_group_title, anzsic_class_code, anzsic_class_title
+  ) %>%
+  dplyr::as_tibble()
 
 # Finalise data frame; noting that we are avoiding the nfd complication for now
-anzsic2006 <- anzsic_raw %>%
+anzsic2006 <- anzsic_2006_final %>%
   arrange(anzsic_division_code,
           anzsic_subdivision_code,
           anzsic_group_code,

pkgdown/_pkgdown.yml

@@ -27,13 +27,15 @@ reference:
       - starts_with("asced")
       - starts_with("asc")
       - auholidays
+      - school_terms
   - title: "Importing ABS Data"
     desc: >
       Functions for retrieving ABS data
     contents:
       - read_absmap
       - get_seifa
       - get_seifa_index_sheet
+      - read_correspondence_tbl
   - title: "Helper functions"
     desc: >
       Functions for cleaning data and working with datasets