geom_text with stats - how to account for fill aesthetic

[leoncio]

Hi there,
the problem is geom_text places text one over the other when fill splits the data into subgroups, e.g.:

starwars |> 
    ggplot()+
     aes(x = gender, y = height, fill = sex)+
     geom_boxplot(width=0.3, size=1, fatten=2, colour="grey50", ) +
    stat_summary(fun = mean, color = "black", position = position_dodge(0.3), geom = "point", shape = 18, size = 5, show.legend = FALSE) +
     geom_text(aes(label=after_stat(count)), stat='count', colour="red", size=4, y = 250) +
     theme_classic()

How to set text below the appropriate boxplots?

[mduvekot]

In stead of geom_text(), you could use this:

  stat_summary(
    fun.data = function(grouped_height) {
      data.frame(
        y = mean(grouped_height),
        label = length(grouped_height)
      )
    },
    geom = "text",
    colour = "red",
    size = 4,
    vjust = -0.0
  )

[leoncio]

@mduvekot

stat_summary(
    fun.data = function(grouped_height) {
      data.frame(
        y = mean(grouped_height),
        label = length(grouped_height)
      )
    },
    geom = "text",
    colour = "red",
    size = 4,
    vjust = -0.0
  )

The problem is, the values, when set to show below boxplots (e.g. y = 1) still are overlayed one over the other...

[mduvekot]

This works for me:

library(ggplot2)
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

starwars |> 
  select(gender, height, sex) |> 
  filter(if_all(everything(), ~ !is.na(.))) |> 
  ggplot()+
  aes(x = gender, y = height, fill = sex)+
  geom_boxplot(
    width=0.3, 
    size=1, 
    median.linewidth=2, 
    colour="grey50") +
  stat_summary( 
    fun = mean, 
    color = "black", 
    position = position_dodge(0.3), 
    geom = "point", 
    shape = 18, 
    size = 5, 
    show.legend = FALSE
  ) +
  stat_summary(
    fun.data = function(grouped_height) {
      data.frame(
        y = mean(grouped_height),
        label = length(grouped_height)
      )
    },
    geom = "text",
    colour = "red",
    size = 4,
    vjust = -0.0
  ) + 
  theme_classic()

^{Created on 2025-08-07 with reprex v2.1.1}

[leoncio]

I have plotted this, but still numbers are on one vertical axis - when you try to put them at the bottom of the fig they are on top of each other (I don't want them at the mean or median's height ).
I would need them below the relevant boxplots, not aside ;).

[mduvekot]

library(dplyr, quietly = TRUE)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
library(ggplot2)

starwars |>
  select(gender, height, sex) |> 
  filter(if_all(everything(), ~ !is.na(.))) |> 
  ggplot(aes(x = gender, y = height, fill = sex)) +
  geom_boxplot(width = 0.3, size = 1, fatten = 2, colour = "grey50") +
  stat_summary(
    fun = mean,
    geom = "point",
    shape = 18,
    size = 5,
    color = "black",
    # position = position_dodge(0.3),
    show.legend = FALSE
  ) +
  stat_summary(
    fun.data = function(grouped_height) {
      data.frame(
        y = quantile(grouped_height, 0.25, na.rm = TRUE),
        label = length(grouped_height)
      )
    },
    position = position_dodge(.3),
    geom = "text",
    colour = "red",
    size = 4,
    vjust = 1.5
  )
#> Warning: The `fatten` argument of `geom_boxplot()` is deprecated as of ggplot2 4.0.0.
#> ℹ Please use the `median.linewidth` argument instead.
#> This warning is displayed once every 8 hours.
#> Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
#> generated.

^{Created on 2025-08-08 with reprex v2.1.1}

[leoncio]

I've done slight corrections and now it works like it should, thanks!

I am posting it below - will have to think/check why it had to be a lambda function there

library(tidyverse)

starwars |>
     select(gender, height, sex) |> 
     filter(if_all(everything(), ~ !is.na(.))) |> 
     ggplot(aes(x = gender, y = height, fill = sex)) +
     geom_boxplot(width = 0.3, size = 1, fatten = 2, colour = "grey50") +
     stat_summary(
         fun = mean,
         geom = "point",
         shape = 18,
         size = 5,
         color = "black",
         position = position_dodge(0.3),
         show.legend = FALSE
     ) +
     stat_summary(
         fun.data = function(grouped_height) {
             data.frame(
                 y = 1,
                 label = length(grouped_height)
             )
         },
         position = position_dodge(.3),
         geom = "text",
         colour = "red",
         size = 4,
         vjust = 1.5
     )

EDIT: Sorry -meant to mark "solution" under your post @mduvekot, now cannot unmark this...