Add column and row percentage to contingency table

Question

I'm trying to add the column and row percentage (simultaneously) to a contingency table.

I tried something like this:

library(tidyverse)
library(janitor)

mtcars %>% 
  tabyl(gear, cyl) %>% 
  adorn_totals("row", name = "Column Total") %>% 
  adorn_totals("col") %>% 
  adorn_percentages("col") %>%
  adorn_pct_formatting() %>%
  adorn_ns()

         gear           4          6           8       Total
            3   9.1%  (1)  28.6% (2)  85.7% (12)  46.9% (15)
            4  72.7%  (8)  57.1% (4)   0.0%  (0)  37.5% (12)
            5  18.2%  (2)  14.3% (1)  14.3%  (2)  15.6%  (5)
 Column Total 100.0% (11) 100.0% (7) 100.0% (14) 100.0% (32)

But I only get the column total (each column sums 100%). However I would like to have the row percentage at the same time. In other words, each cell would contain the N, the column percentage and the row percentage, although (eventually) I could sacrifice the N.

Answer 1

janitor is designed for quite simple tables, so (unless I'm not mistaken) in order to create this we have to use other tools

totals <- mtcars %>% 
  tabyl(gear, cyl) %>% 
  adorn_totals("row", name = "Row Total") %>% 
  adorn_totals("col", name = "Column Total")

rows <- totals %>%
  adorn_percentages("row") %>%
  adorn_pct_formatting()

cols <- totals %>%
  adorn_percentages("col") %>%
  adorn_pct_formatting()

totals %>%
    mutate(across(-gear, ~paste0(., " (", rows[[cur_column()]], ", ", cols[[cur_column()]], ")")))

      gear                  4                 6                  8
         3     1 (6.7%, 9.1%)  2 (13.3%, 28.6%)  12 (80.0%, 85.7%)
         4   8 (66.7%, 72.7%)  4 (33.3%, 57.1%)     0 (0.0%, 0.0%)
         5   2 (40.0%, 18.2%)  1 (20.0%, 14.3%)   2 (40.0%, 14.3%)
 Row Total 11 (34.4%, 100.0%) 7 (21.9%, 100.0%) 14 (43.8%, 100.0%)
        Column Total
  15 (100.0%, 46.9%)
  12 (100.0%, 37.5%)
   5 (100.0%, 15.6%)
 32 (100.0%, 100.0%)

I wasn't sure what format you wanted, so I went with n, row%, col%, but you can change it if you like (just adjust the last mutate call)