CODEWITHSUNDEEP
rggplot2
Pengin
Pengin

Reputation: 4772

ggplot2 legend for stat_summary

How can I create a legend informing that the red cross is the mean?

ggplot(results, aes(x=factor, y=proportionPositive)) +
geom_boxplot() +
stat_summary(fun.data = "mean_cl_normal", colour = "red", shape=4)

enter image description here

Upvotes: 19

Views: 13300

Answers (3)

asenga
asenga

Reputation: 21

for me option "show.legend=TRUE" simply did work out:

ggplot(aes(x=stimulus, y=EPN, fill=strategy))+ 
  stat_summary(fun.data=mean_se, show.legend=TRUE, geom="bar", position="dodge", colour="black", linetype="solid", size=0.3)

Upvotes: 1

Tunn
Tunn

Reputation: 1536

To make it appear like a default legend (borrowing from @Andrie code):

ggplot(results, aes(x=factor, y=proportionPositive)) +
      geom_boxplot() +
      stat_summary(fun.data = "mean_cl_normal", 
              aes(shape=""), # Leave empty
              colour = "red",
              geom="point") +
      scale_shape_manual("mean", values= "") # Will show mean on top of the line

Upvotes: 0

Andrie
Andrie

Reputation: 179398

Here is one way of doing it:

  1. Map an aesthetic to a shape, i.e. aes(shape="mean")
  2. Create a manual shape scale, i.e. scale_shape_manual()
# Create dummy data
results <- data.frame(
  factor=factor(rep(1:10, 100)), 
  proportionPositive=rnorm(1000))

# Plot results
ggplot(results, aes(x=factor, y=proportionPositive)) +
      geom_boxplot() +
      stat_summary(fun.data = "mean_cl_normal", 
              aes(shape="mean"), 
              colour = "red",
              geom="point") +
      scale_shape_manual("", values=c("mean"="x"))

enter image description here

Upvotes: 28

Related Questions