CODEWITHSUNDEEP
rggplot2likert
Homer Jay Simpson
Homer Jay Simpson

Reputation: 1278

sort the gglikert chart based on the combination percentages in R

I came along the gglikert() function in R from the ggsstats library here. They used sampling data and create a data frame called df. I am using the same as given in the link.(the df not the df_dk). Ok, if i run in R the following code :

p = ggstats::gglikert(df, sort = "descending",
                  sort_method = c("prop"))
p

i receive a plot :

enter image description here

which is sorted in an ascending order and the "prop" sort according to the proportion of answers higher than the centered level. I want the plot to be sorted according to very left proportions that are the combination (sum) of the two lower levels. (i.e Respectively the percentages on the very right in the sum of the two lower categories. )

how can i achieve this in R ?

Upvotes: 1

Views: 144

Answers (2)

Joseph Larmarange
Joseph Larmarange

Reputation: 116

If you look at the documentation https://larmarange.github.io/ggstats/reference/gglikert.html there is an argument sort_prop_include_center that could be set up to TRUE in order to include half of the centered level when sorting data.

Similarly, the argument totals_include_center allows you to include half of the centered level into the left and the right totals

Upvotes: 2

stefan
stefan

Reputation: 123758

Here is one possible option which uses reorder and an ifelse to reorder the variable mapped on y using the counts (aka the sum) of Strictly disagree and disagree answers. Under the hood ggstats::gglikert reshape the data to long where the question id's are stored in a column named .question and the answers in a column named .answer:

library(ggplot2)
library(ggstats)

ggstats::gglikert(df) +
  aes(y = reorder(.question,
    ifelse(
      .answer %in% c("Strongly disagree", "Disagree"),
      1, 0
    ),
    FUN = sum
  ))

DATA

likert_levels <- c(
  "Strongly disagree",
  "Disagree",
  "Neither agree nor disagree",
  "Agree",
  "Strongly agree"
)
set.seed(42)

library(dplyr, warn = FALSE)

df <-
  tibble(
    q1 = sample(likert_levels, 150, replace = TRUE),
    q2 = sample(likert_levels, 150, replace = TRUE, prob = 5:1),
    q3 = sample(likert_levels, 150, replace = TRUE, prob = 1:5),
    q4 = sample(likert_levels, 150, replace = TRUE, prob = 1:5),
    q5 = sample(c(likert_levels, NA), 150, replace = TRUE),
    q6 = sample(likert_levels, 150, replace = TRUE, prob = c(1, 0, 1, 1, 0))
  ) %>%
  mutate(across(everything(), ~ factor(.x, levels = likert_levels)))

Upvotes: 1

Related Questions