CODEWITHSUNDEEP
rggplot2ggpubr
maycca
maycca

Reputation: 4090

R ggpubr: move legend title above the legend keys

I am struggling with the legend position within ggpubr. I know that I can modify the legend position p.e. by ggpar(legend = "bottom"). But, how to place legend title above the legend keys?

In ggplot2, it seems that guide_legend(title.position = "top") should do this, but how to make it work with ggpubr? (thanks to @c06n and link here: https://ggplot2.tidyverse.org/reference/guide_legend.html). I would like to combine the ggplot2 and ggpubr but I don't know exactly why.

enter image description here

Dummy example:

df2 <- data.frame(supp=rep(c("VC", "OJ"), each=3),
                  dose=rep(c("D0.5", "D1", "D2"),2),
                  len=c(6.8, 15, 33, 4.2, 10, 29.5))

# Plot "len" by "dose" and
# Change line types and point shapes by a second groups: "supp"
p<- ggline(df2, "dose", "len",
       linetype = "supp", shape = "supp",
       color = "supp", palette = c("#00AFBB", "#E7B800"))

# Seems not working
ggpar(p,
  linetype = guide_legend(title = "My title", title.position = "top"))

Upvotes: 3

Views: 4052

Answers (2)

c06n
c06n

Reputation: 176

I do not know ggpubr, it seems like a wrapper for ggplot2. If that is true, and/or it has implemented the same functionality as ggplot2, you should be able to achieve that via legend guide, see here. I think it is title.position you need to adjust.

If you cannot achieve that with ggpubr, I suggest using ggplot2 instead, perhaps adjusting the theme to your needs.

Edit. @markus has the answer, so to put it all in one place:

p <- ggline(df2, "dose", "len",
           linetype = "supp", 
           shape = "supp",
           color = "supp", 
           palette = c("#00AFBB", "#E7B800")) +
  guides(color = guide_legend(title.position = "top", 
                              title.hjust = 0.5))

Interesting question though from @Bruno Pinheiro as to why only color (and fill) will work, but not shape or linetype. This would be relevant if the plot needs to be monochrome. All three should work equally well as grouping factors.

Does anyone have an idea?

Upvotes: 4

Bruno Pinheiro
Bruno Pinheiro

Reputation: 1014

These options are working here.

With ggpubr:

df2 <- data.frame(supp=rep(c("VC", "OJ"), each=3),
                  dose=rep(c("D0.5", "D1", "D2"),2),
                  len=c(6.8, 15, 33, 4.2, 10, 29.5))

library(ggpubr)
p <- ggline(df2, "dose", "len",
       linetype = "supp", shape = "supp",
       color = "supp", palette = c("#00AFBB", "#E7B800"))

p + guides(colour = guide_legend(title.position = "top"))

I don't know why, but if i set the guide for another aesthetic the legend title position do not change:

p + guides(shape = guide_legend(title.position = "top"))
p + guides(linetype = guide_legend(title.position = "top"))

It's a good question to be made for ggplot2 experts.

And here is the same working in ggplot2:

library(ggplot2)
ggplot(df2, aes(x = dose, y = len, colour = supp)) +
  geom_line(aes(group = supp, linetype = supp)) +
  geom_point(aes(shape = supp)) +
  scale_colour_manual(values = c("#00AFBB", "#E7B800")) +
  theme_classic() +
  theme(legend.position = "top") +
  guides(colour = guide_legend(title.position = "top"))

Upvotes: 1

Related Questions