CODEWITHSUNDEEP
rggplot2
Nicholas
Nicholas

Reputation: 1436

Adding vertical lines through median values of cumulative distributions (ggplot2)

I'm using ggplot2 to plot two cumulative distributions on a single plot. This is straightforward using the example from ?stat_ecdf. My difficulty is in adding vertical lines through the median value of each distribution.

It's easy enough to do this with a single distribution:

df <- data.frame(x = c(rnorm(100, 5, 10), rnorm(200, 0, 10)),
  g = as.factor(c(rep(1, 100), rep(2, 200))))

ggplot(df, aes(x)) +
  stat_ecdf() + 
  geom_vline(aes(xintercept = median(x)))

But I can't figure out an easy way to add vertical lines for multiple distributions. I've tried the following without success:

ggplot(df, aes(x, colour = g)) +
  stat_ecdf() +   
  geom_vline(aes(xintercept = median(x), colour = g))

I can sort of get the desired result by assembling the plot in stages:

p <- ggplot(df[df$g == 1, ], aes(x)) +
  stat_ecdf() + 
  geom_vline(aes(xintercept = median(x)))

p + 
  stat_ecdf(data = df[df$g == 2, ]) +
  geom_vline(data = df[df$g == 2, ], aes(xintercept = median(x)))

But this seems like an untidy way of doing it and also leaves me to set the different line colours by hand.

Surely there's a better way?

Upvotes: 3

Views: 6684

Answers (1)

rmbaughman
rmbaughman

Reputation: 921

Try this:

ggplot(df, aes(x, colour = g)) +
    stat_ecdf() +   
    geom_vline(aes(xintercept = median(x[g==1]), color=g[g==1])) + 
    geom_vline(aes(xintercept = median(x[g==2]), color=g[g==2]))

By adding a second geom_vline argument and specifying which group each line and color is referring to, you can get two lines with the corresponding colors. Hope that helps!

Upvotes: 5

Related Questions