Is there a way for adding labels with number of observations per value in violin plot in ggplot?

Question

Image you want to create a violin plot and have data like this:

set.seed(123)
Bodytype <- sample(LETTERS[1:3], 500, replace = T)
Weight <- rnorm(500,40,1)    
df <- data.frame(Bodytype, Weight)
  ggplot(data = df, 
         aes(x = Bodytype, y = Weight, fill = Bodytype)) +
  geom_violin(scale = "count", trim = F, adjust = 0.75) +
  scale_y_continuous(breaks = seq(34, 46, 1)) +
  theme_gray() 

Now I would like to add text label or something for each bodytype at each kg level to see how many observations in each bodytype category that weigh 36 kg, 37kg, etc. Is there are way for achieving this, or am I better off using another plot?

Answer 1

This can be done in many ways, here is one:

library(dplyr)
library(ggplot2)
summ <- df %>%
  group_by(Bodytype) %>%
  summarize(n = n(), Weight = mean(Weight))
ggplot(data = df, aes(x = Bodytype, y = Weight, fill = Bodytype)) +
  geom_violin(scale = "count", trim = F, adjust = 0.75) +
  scale_y_continuous(breaks = seq(34, 46, 1)) +
  theme_gray() +
  geom_text(aes(label = n), data = summ)

---

Okay, so you want multiple weight counts:

weightcounts <- df %>%
  mutate(Weight = as.integer(round(Weight, 0))) %>%
  group_by(Bodytype, Weight) %>%
  count()
ggplot(data = df, aes(x = Bodytype, y = Weight, fill = Bodytype)) +
  geom_violin(scale = "count", trim = F, adjust = 0.75) +
  scale_y_continuous(breaks = seq(34, 46, 1)) +
  theme_gray() +
  geom_text(aes(label = n), data = weightcounts)

Either way, the premise is that you can generate a summary frame with the associated labels you need, and then add geom_text (or geom_label) with the new dataset as an argument.

Answer 2

Another way computing labels outside in base R:

set.seed(123)
Bodytype <- sample(LETTERS[1:3], 500, replace = T)
Weight <- rnorm(500,40,1)    
df <- data.frame(Bodytype, Weight)
#Labels
df$i <- 1
labs <- aggregate(i~Bodytype,df,sum)
labs$Weight<-NA
#Plot
ggplot(data = df, 
       aes(x = Bodytype, y = Weight, fill = Bodytype)) +
  geom_violin(scale = "count", trim = F, adjust = 0.75) +
  geom_text(data=labs,aes(x=Bodytype,y=45,label=i))
  scale_y_continuous(breaks = seq(34, 46, 1)) +
  theme_gray() 

Output: