Adding percentage labels to a bar chart in ggplot2

Question

How can I use geom_text to add percentage labels on top of each bar in ggplot2? I know there are several similar questions which are already answered. But they either use only 1 categorical variable or compute the percentages before plotting.
I have following plot:

ggplot(data = mtcars)+
  geom_bar(aes(x = factor(cyl), 
               y = (..count..)/sum(..count..)*100,
               fill = factor(gear)),
           position = "dodge")  

Now I want to add the percentage labels on the top. If I use y = (..count..)/sum(..count..)*100 in geom_text, it says Error in eval(expr, envir, enclos) : object 'count' not found.

Answer 1

It's easiest to calculate the quantities you need beforehand, outside of ggplot, as it's hard to track what ggplot calculates and where those quantities are stored and available.

First, summarize your data:

library(dplyr)
library(ggplot2)

mtcars %>% 
    count(cyl = factor(cyl), gear = factor(gear)) %>% 
    mutate(pct = prop.table(n))
#> # A tibble: 8 x 4
#>   cyl   gear      n    pct
#>   <fct> <fct> <int>  <dbl>
#> 1 4     3         1 0.0312
#> 2 4     4         8 0.25  
#> 3 4     5         2 0.0625
#> 4 6     3         2 0.0625
#> 5 6     4         4 0.125 
#> 6 6     5         1 0.0312
#> 7 8     3        12 0.375 
#> 8 8     5         2 0.0625

Save that if you like, or pipe directly into ggplot:

mtcars %>% 
    count(cyl = factor(cyl), gear = factor(gear)) %>% 
    mutate(pct = prop.table(n)) %>% 
    ggplot(aes(x = cyl, y = pct, fill = gear, label = scales::percent(pct))) + 
    geom_col(position = 'dodge') + 
    geom_text(position = position_dodge(width = .9),    # move to center of bars
              vjust = -0.5,    # nudge above top of bar
              size = 3) + 
    scale_y_continuous(labels = scales::percent)

If you really want to keep it all internal to ggplot, you can use geom_text with stat = 'count' (or stat_count with geom = "text", if you prefer):

ggplot(data = mtcars, aes(x = factor(cyl), 
                          y = prop.table(stat(count)), 
                          fill = factor(gear), 
                          label = scales::percent(prop.table(stat(count))))) +
    geom_bar(position = "dodge") + 
    geom_text(stat = 'count',
              position = position_dodge(.9), 
              vjust = -0.5, 
              size = 3) + 
    scale_y_continuous(labels = scales::percent) + 
    labs(x = 'cyl', y = 'pct', fill = 'gear')

which plots exactly the same thing.