Counts and proportions within nested dataframes - group-by

Question

I would like to count and calculate proportions within several subdataframes. I used to do this with some ugly code, using lapply and creating and looping over many objects. With the purrr package this should be more straightforward, however I didn't manage to do it so far.
Illustration with the standard R dataset "mtcars":
The following code is a standard dplyr solution (which works):
mtcars%>% group_by(vs, am) %>% summarise(n = n()) %>% mutate(freq = prop.table(n))
Thus, I get the proportions and counts within the levels of "vs" (0 and 1). However, I want to calculate the grouped counts and proportions within several subgroups of "carb". So with standard dplyr this would look like that (works as well):
mtcars %>% filter(carb == 1) %>% summarise(n = n()) %>% mutate(freq = prop.table(n))
mtcars %>% filter(carb == 2) %>% summarise(n = n()) %>% mutate(freq = prop.table(n))
etc.
This works but is cumbersome.
With Purrr it should look something like that:
mtcars %>% group_by(carb) %>% nest() %>% mutate(n = map(data, count))
Here however, the grouping group_by(vs, am) is lost. How can I introduce the grouping here?
Many thanks in advance!

Answer 1

Is this the desired ouput?
library(tidyverse)
unique(mtcars$carb) %>% map_dfr(~ mtcars %>%
group_by(vs, am) %>%
filter(carb == .x) %>%
summarise(carb = .x, n = n(),.groups = 'drop') %>%
group_by(vs) %>%
mutate(freq = prop.table(n)))
#> # A tibble: 12 × 5
#> # Groups: vs [2]
#> vs am carb n freq
#> <dbl> <dbl> <dbl> <int> <dbl>
#> 1 0 0 4 5 0.625
#> 2 0 1 4 3 0.375
#> 3 1 0 4 2 1
#> 4 1 0 1 3 0.429
#> 5 1 1 1 4 0.571
#> 6 0 0 2 4 0.8
#> 7 0 1 2 1 0.2
#> 8 1 0 2 2 0.4
#> 9 1 1 2 3 0.6
#> 10 0 0 3 3 1
#> 11 0 1 6 1 1
#> 12 0 1 8 1 1
Created on 2022-01-06 by the reprex package (v2.0.1)