Inconsistent labels between Indonesian & English datasets

[dzulfiqarfr]

Hi, folks. If you request a dataset in English either using bpsr or Statistics Indonesia's (BPS) website, you may want to check it against the Indonesian version to see whether the labels are consistent. While I haven't tried to check all datasets, I've found that the food, beverage and tobacco consumer price index (CPI) dataset has inconsistent labels.

The Indonesian dataset, for example, labels the derived variable ID "1553" as nonalcoholic beverages.

library(conflicted)
library(bpsr)
library(dplyr)
conflict_prefer("filter", "dplyr")

cpi_ind <- bps_get_dataset("1905", keep = "derived_var_id")

filter(cpi_ind, derived_var_id == "1553")
# A tibble: 3,094 × 6
   vertical_var  derived_var_id derived_var                    year period      var
   <chr>         <chr>          <chr>                         <int> <chr>     <dbl>
 1 KOTA MEULABOH 1553           Minuman yang Tidak Beralkohol  2020 Januari    103.
 2 KOTA MEULABOH 1553           Minuman yang Tidak Beralkohol  2020 Februari   103.
 3 KOTA MEULABOH 1553           Minuman yang Tidak Beralkohol  2020 Maret      103.
 4 KOTA MEULABOH 1553           Minuman yang Tidak Beralkohol  2020 April      103.
 5 KOTA MEULABOH 1553           Minuman yang Tidak Beralkohol  2020 Mei        103.
 6 KOTA MEULABOH 1553           Minuman yang Tidak Beralkohol  2020 Juni       103.
 7 KOTA MEULABOH 1553           Minuman yang Tidak Beralkohol  2020 Juli       103.
 8 KOTA MEULABOH 1553           Minuman yang Tidak Beralkohol  2020 Agustus    103.
 9 KOTA MEULABOH 1553           Minuman yang Tidak Beralkohol  2020 September  103.
10 KOTA MEULABOH 1553           Minuman yang Tidak Beralkohol  2020 Oktober    103.
# … with 3,084 more rows
# ℹ Use `print(n = ...)` to see more rows
# ℹ Read the metadata with `bps_metadata()`

But the English dataset labels the same derived variable ID as alcoholic beverages.

cpi_eng <- bps_get_dataset("1905", lang = "eng", keep = "derived_var_id")

filter(cpi_eng, derived_var_id == "1553")
# A tibble: 3,094 × 6
   vertical_var  derived_var_id derived_var          year period      var
   <chr>         <chr>          <chr>               <int> <chr>     <dbl>
 1 KOTA MEULABOH 1553           Alcoholic Beverages  2020 January    103.
 2 KOTA MEULABOH 1553           Alcoholic Beverages  2020 February   103.
 3 KOTA MEULABOH 1553           Alcoholic Beverages  2020 March      103.
 4 KOTA MEULABOH 1553           Alcoholic Beverages  2020 April      103.
 5 KOTA MEULABOH 1553           Alcoholic Beverages  2020 May        103.
 6 KOTA MEULABOH 1553           Alcoholic Beverages  2020 June       103.
 7 KOTA MEULABOH 1553           Alcoholic Beverages  2020 July       103.
 8 KOTA MEULABOH 1553           Alcoholic Beverages  2020 August     103.
 9 KOTA MEULABOH 1553           Alcoholic Beverages  2020 September  103.
10 KOTA MEULABOH 1553           Alcoholic Beverages  2020 October    103.
# … with 3,084 more rows
# ℹ Use `print(n = ...)` to see more rows
# ℹ Read the metadata with `bps_metadata()`

You can also see this inconsistency in labelling on the agency's website as well. This implies that the inconsistency emerges not as a result of incorrect parsing by bpsr. I've tried to ensure that both Indonesian and English datasets are parsed correctly by formally testing this.

I have reported this issue to the contact listed on BPS API's website. Let's hope they'll be able to sort this out. 🤞🏽

[dzulfiqarfr]

BPS has responded and confirmed there was an incorrect translation in the food, beverage and tobacco CPI dataset. The team has solved this issue.

library(conflicted)
library(bpsr)
library(dplyr)
conflict_prefer("filter", "dplyr")

cpi_eng <- bps_get_dataset("1905", lang = "eng", keep = "derived_var_id")

filter(cpi_eng, derived_var_id == "1553")
# A tibble: 3,185 × 6
   vertical_var  derived_var_id derived_var              year period      var
   <chr>         <chr>          <chr>                   <int> <chr>     <dbl>
 1 KOTA MEULABOH 1553           Non-Alcoholic Beverages  2020 January    103.
 2 KOTA MEULABOH 1553           Non-Alcoholic Beverages  2020 February   103.
 3 KOTA MEULABOH 1553           Non-Alcoholic Beverages  2020 March      103.
 4 KOTA MEULABOH 1553           Non-Alcoholic Beverages  2020 April      103.
 5 KOTA MEULABOH 1553           Non-Alcoholic Beverages  2020 May        103.
 6 KOTA MEULABOH 1553           Non-Alcoholic Beverages  2020 June       103.
 7 KOTA MEULABOH 1553           Non-Alcoholic Beverages  2020 July       103.
 8 KOTA MEULABOH 1553           Non-Alcoholic Beverages  2020 August     103.
 9 KOTA MEULABOH 1553           Non-Alcoholic Beverages  2020 September  103.
10 KOTA MEULABOH 1553           Non-Alcoholic Beverages  2020 October    103.
# … with 3,175 more rows
# ℹ Use `print(n = ...)` to see more rows
# ℹ Read the metadata with `bps_metadata()`