How to tidy a deeply nested json-File in R? [closed] - r

Question

Closed. This question needs details or clarity. It is not currently accepting answers.
Want to improve this question? Add details and clarify the problem by editing this post.
Closed 10 months ago.
Improve this question
There is a json-File I would like to analyze:
https://dam-api.bfs.admin.ch/hub/api/dam/assets/21364129/master
It has data about four different referendums (all from the same date), all the results for each region (kanton) and on community-level (gemeinde). I tried to make a tidy df out of this, but I don't even understand the exact syntax of the unnest-command (cols?).
After a lot of try and error, I can get the data for a specific kanton, for example like this:
flatten(df$schweiz$vorlagen$kantone[[4]]$gemeinden[[13]])
But there must be some way to get this data for all 26 cantons, and of course with the information to which one of the four referendum the data belongs. I guess that some tools from the purrr-package might be helpful, but all the tutorials i read so far have been a mistery to me.
So, in short: Is there a way to get this file tidy without hours of manual work?

Answer 1

Converting such a large and complex structure into a data frame requires specific knowledge of what you want your final output to be, and you really need to tailor your code to do it. In this case, I assume you want a list containing 4 data frames (one for each referendum), where each row gives the kantone, the gemeinden, and the various fields indicating the result from each gemeinden. This is complex, but something like the following should work:
url <- "https://dam-api.bfs.admin.ch/hub/api/dam/assets/21364129/master"
json <- jsonlite::read_json(url)
result <- lapply(json$schweiz$vorlagen, function(x) {
dplyr::as_tibble(do.call(rbind, lapply(x$kantone, function(y) {
k <- y$geoLevelname
do.call(rbind, lapply(y$gemeinden, function(z) {
cbind(data.frame(kantone = k, gemeinden = z$geoLevelname),
as.data.frame(z$resultat))
}))
})))
})
So your result looks like this:
result
#> [[1]]
#> # A tibble: 2,156 x 10
#> kantone gemeinden gebietAusgezaeh~ jaStimmenInProz~ jaStimmenAbsolut
#> <chr> <chr> <lgl> <dbl> <int>
#> 1 Zürich Aeugst am Albis TRUE 25.4 194
#> 2 Zürich Affoltern am Albis TRUE 20.4 661
#> 3 Zürich Bonstetten TRUE 17.9 345
#> 4 Zürich Hausen am Albis TRUE 20.9 286
#> 5 Zürich Hedingen TRUE 19.3 269
#> 6 Zürich Kappel am Albis TRUE 24.1 93
#> 7 Zürich Knonau TRUE 21.6 161
#> 8 Zürich Maschwanden TRUE 21.1 49
#> 9 Zürich Mettmenstetten TRUE 19.3 368
#> 10 Zürich Obfelden TRUE 21.5 325
#> # ... with 2,146 more rows, and 5 more variables: neinStimmenAbsolut <int>,
#> # stimmbeteiligungInProzent <dbl>, eingelegteStimmzettel <int>,
#> # anzahlStimmberechtigte <int>, gueltigeStimmen <int>
#>
#> [[2]]
#> # A tibble: 2,156 x 10
#> kantone gemeinden gebietAusgezaeh~ jaStimmenInProz~ jaStimmenAbsolut
#> <chr> <chr> <lgl> <dbl> <int>
#> 1 Zürich Aeugst am Albis TRUE 60.3 470
#> 2 Zürich Affoltern am Albis TRUE 61.2 2014
#> 3 Zürich Bonstetten TRUE 59.8 1166
#> 4 Zürich Hausen am Albis TRUE 57.7 796
#> 5 Zürich Hedingen TRUE 58.5 826
#> 6 Zürich Kappel am Albis TRUE 49.6 192
#> 7 Zürich Knonau TRUE 56.8 431
#> 8 Zürich Maschwanden TRUE 52.6 123
#> 9 Zürich Mettmenstetten TRUE 58.3 1119
#> 10 Zürich Obfelden TRUE 57.2 868
#> # ... with 2,146 more rows, and 5 more variables: neinStimmenAbsolut <int>,
#> # stimmbeteiligungInProzent <dbl>, eingelegteStimmzettel <int>,
#> # anzahlStimmberechtigte <int>, gueltigeStimmen <int>
#>
#> [[3]]
#> # A tibble: 2,156 x 10
#> kantone gemeinden gebietAusgezaeh~ jaStimmenInProz~ jaStimmenAbsolut
#> <chr> <chr> <lgl> <dbl> <int>
#> 1 Zürich Aeugst am Albis TRUE 39.6 301
#> 2 Zürich Affoltern am Albis TRUE 35.7 1150
#> 3 Zürich Bonstetten TRUE 36.2 697
#> 4 Zürich Hausen am Albis TRUE 39.5 531
#> 5 Zürich Hedingen TRUE 37.1 509
#> 6 Zürich Kappel am Albis TRUE 42.9 162
#> 7 Zürich Knonau TRUE 36.7 273
#> 8 Zürich Maschwanden TRUE 37.1 86
#> 9 Zürich Mettmenstetten TRUE 39.2 742
#> 10 Zürich Obfelden TRUE 34.4 513
#> # ... with 2,146 more rows, and 5 more variables: neinStimmenAbsolut <int>,
#> # stimmbeteiligungInProzent <dbl>, eingelegteStimmzettel <int>,
#> # anzahlStimmberechtigte <int>, gueltigeStimmen <int>
#>
#> [[4]]
#> # A tibble: 2,156 x 10
#> kantone gemeinden gebietAusgezaeh~ jaStimmenInProz~ jaStimmenAbsolut
#> <chr> <chr> <lgl> <dbl> <int>
#> 1 Zürich Aeugst am Albis TRUE 39.0 300
#> 2 Zürich Affoltern am Albis TRUE 40.8 1321
#> 3 Zürich Bonstetten TRUE 42.8 826
#> 4 Zürich Hausen am Albis TRUE 41.6 570
#> 5 Zürich Hedingen TRUE 44.1 615
#> 6 Zürich Kappel am Albis TRUE 35.5 136
#> 7 Zürich Knonau TRUE 38.6 290
#> 8 Zürich Maschwanden TRUE 42.7 100
#> 9 Zürich Mettmenstetten TRUE 39.6 757
#> 10 Zürich Obfelden TRUE 40.5 610
#> # ... with 2,146 more rows, and 5 more variables: neinStimmenAbsolut <int>,
#> # stimmbeteiligungInProzent <dbl>, eingelegteStimmzettel <int>,
#> # anzahlStimmberechtigte <int>, gueltigeStimmen <int>
If you want them all in one big data frame, you can add a number or name column to indicate which referendum you are referring to, then bind all the data frames together.
Created on 2022-04-16 by the reprex package (v2.0.1)