Search code examples
rdplyriterationflextablegtsummary

How to convert a tibble in as many tables as categorical levels variables

I have the following tibble

structure(list(blockLabel = structure(c(1L, 2L, 3L, 4L, 5L, 1L, 
2L, 3L, 4L, 5L, 1L, 2L, 3L, 4L, 5L, 1L, 2L, 3L, 4L, 5L, 1L, 2L, 
3L, 4L, 5L, 1L, 2L, 3L, 4L, 5L), .Label = c("auditory_only", 
"bimodal_focus_auditory", "bimodal_focus_visual", "divided", 
"visual_only"), class = "factor"), trial_resp.corr = structure(c(1L, 
1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 
2L, 2L, 2L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L), .Label = c("0", 
"1"), class = "factor"), participant = structure(c(1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L), .Label = c("pilot01", 
"pilot02", "pilot03"), class = "factor"), Freq = c(0L, 1L, 3L, 
74L, 0L, 12L, 71L, 69L, 70L, 12L, 0L, 1L, 2L, 77L, 11L, 12L, 
71L, 70L, 67L, 1L, 1L, 1L, 3L, 75L, 0L, 11L, 71L, 69L, 69L, 12L
), tc = c(12, 72, 72, 144, 12, 12, 72, 72, 144, 12, 12, 72, 72, 
144, 12, 12, 72, 72, 144, 12, 12, 72, 72, 144, 12, 12, 72, 72, 
144, 12), freq = c(0, 1.38888888888889, 4.16666666666667, 51.3888888888889, 
0, 100, 98.6111111111111, 95.8333333333333, 48.6111111111111, 
100, 0, 1.38888888888889, 2.77777777777778, 53.4722222222222, 
91.6666666666667, 100, 98.6111111111111, 97.2222222222222, 46.5277777777778, 
8.33333333333333, 8.33333333333333, 1.38888888888889, 4.16666666666667, 
52.0833333333333, 0, 91.6666666666667, 98.6111111111111, 95.8333333333333, 
47.9166666666667, 100)), class = c("tbl_df", "tbl", "data.frame"
), row.names = c(NA, -30L))

I would like to create three different tables according to the levels of participant variable. More or less the final result should be like the following one:

enter image description here

I have started scripting the following code (since I am looking for to do this via a dplyr, apply family, loops or map function)

  list %>%  as_data_frame() %>%
  select(blockLabel, trial_resp.corr, participant, Freq, freq) %>% 
  map(~ flextable(.x))

But unfortunately, I get the following error code:

Error in flextable(.x) : is.data.frame(data) is not TRUE

I am not an expert in this method thus if you would have something to suggest about fixing the problem and sharing knowledge to achieve the final result, please just let me know (in passing I specify that Correctness corresponds to 1 and Incorrcetness to 0 of trial_resp.corr variable)

Thanks

Solution

  • Here is one method where we modify the 'trial_resp.corr' values to 'Incorrectness', 'Correctness' based on 0, 1 values, then select the columns of interest, split by the 'participant' column to create a list of datasets, loop over the list with map, reshape data from long to wide with pivot_wider and then convert to flextable

    library(dplyr)
library(tidyr)
library(purrr)
library(flextable)
library(stringr)
library(janitor)
lst1 <-  df1 %>% 
  mutate( trial_resp.corr = recode(trial_resp.corr, 
             '0'= 'Incorrectness', '1' = "Correctness")) %>%
  select(blockLabel, trial_resp.corr, participant, Freq,
  `Freq%` = freq) %>%
  split(.$participant) %>% 
  map(~ .x  %>% 
        select(-participant)  %>% 
        pivot_wider(names_from = trial_resp.corr, 
                    values_from = c(Freq, `Freq%`), values_fill = 0, 
            names_glue ="{trial_resp.corr}{str_remove(.value, 'Freq' )}") %>%
        adorn_totals() %>%
        mutate(across(ends_with("%"),
               ~ replace(.x, n(), 
         100 *get(str_remove(cur_column(), 
                               fixed("%")))[n()]/sum(Incorrectness[n()], 
     Correctness[n()])))) %>%
        mutate(across(ends_with("%"),~  round(.x, digits = 2)))%>%
        
        flextable)

    If we want to create objects in the global env

    list2env(lst1, .GlobalEnv)

    -output

    lst1[[1]]

    enter image description here