How to customize a boxplot legend indicating mean, outliers, median, etc?

Question

I have a boxplot and by my supervisor's advice I have to indicate the mean, outliers and median in the legend, like this image:

How can I do this using ggplot2?

library(ggplot2)

A <- 1:20
DF <- data.frame(A)

ggplot(data = DF) +
  geom_boxplot(aes(x = "", y = A))

Answer 1

There is no straightforward way. But you could make a custom legend using another plot:

p <- ggplot(mtcars) +
  geom_boxplot(aes(x = factor(cyl), y = mpg))

d1 <- data.frame(x = 1, y = c(1:1000, 1502))
d2 <- data.frame(
  y = c(boxplot.stats(d1$y)$stats, 1502), 
  x = 1, 
  label = c('min', '1st quartile', 'median', '3rd quartile', 'max', 'outlier')
)
leg <- ggplot(d1, aes(x, y)) + geom_boxplot(width = 0.2) + 
  geom_text(aes(x = 1.15, label = label), d2, hjust = 0) +
  xlim(0.9, 1.5) +
  theme_void() + theme(panel.background = element_rect(fill = 'white', color = 1))

p + annotation_custom(ggplotGrob(leg), xmin = 3, xmax = 3.5, ymin = 25, ymax = 35)