CODEWITHSUNDEEP
rggplot2graphdata-visualizationlegend
Thredolsen
Thredolsen

Reputation: 257

Add a combined legend that accounts for color, shape, and linetype, while keeping the original legends

I have the following df:

df <- data.frame("id" = c("A1", "A1", "A1", "A2", "A2", "A2", "B1", "B1", "B1", "B2", "B2", "B2"),
                 "group" = c("A", "A", "A", "A", "A", "A", "B", "B", "B", "B", "B", "B"),
                 "version" = c("1", "1", "1", "2", "2", "2", "1", "1", "1", "2", "2", "2"),
                 "var_1" = 1:3,
                 "var_2" = 1:12)

I use the following code::

library(ggplot2)

ggplot(data = df, aes(x = var_1, y = var_2, by = "id")) +
  stat_summary(fun = "mean", geom = "line", aes(color = factor(id), linetype = (version))) +
  stat_summary(fun = "mean", geom = "point", aes(color = factor(id), shape = (group))) +
  labs(color = "id", linetype = "version", shape = "group")

And get the following plot:

enter image description here

The problem is that in the 'id' legend, the shapes and linetypes are generic, and don't always match the correct shape/linetype.

What I would like to do is make the legend for 'id' show the correct shape and linetype, while preserving the separate, original legends for 'group' and 'version'.

This means that my goal is for the 'group' and 'version' legends to remain the same, but for the entries in the 'id' legend to have their shape and linetype modified as necessary.

Upvotes: 0

Views: 176

Answers (1)

chemdork123
chemdork123

Reputation: 13843

You can do this through guide_legend. Within that, you can override the default aes() specified via the other ggplot2 commands to match what you want:

p + guides(color=guide_legend(
  override.aes = list(linetype=c(1,3,1,3), shape=c(16,16,17,17))))

enter image description here

Upvotes: 1

Related Questions