ggplot2 : Plot mean with geom_bar

Question

I have the following data frame:

test2 <- data.frame(groups = c(rep("group1",4), rep("group2",4)), 
    X2 = c(rnorm(4), rnorm(4)) , 
    label = c(rep(1,2),rep(2,2),rep(1,2),rep(2,2)))

and I am plotting the bar graphs for each label per group using:

ggplot(test2, aes(label, X2, fill=as.factor(groups))) + 
    geom_bar(position="dodge", stat="identity")

However, I am cannot seem to be able to find a stat="mean" so I can plot the means on each bar graph instead of the identity.

Thanks for any help.

This tutorial provides a good description of how to achieve this: http://www.r-bloggers.com/using-r-barplot-with-ggplot2/ — seaotternerd, May 12 '15 at 06:29
possible duplicate of [Plotting the average values for each level in ggplot2](http://stackoverflow.com/questions/11857935/plotting-the-average-values-for-each-level-in-ggplot2) — seaotternerd, May 12 '15 at 06:32

score 99 · Accepted Answer · edited May 17 '16 at 05:50

99

simply use stat = "summary" and fun.y = "mean"

ggplot(test2) + 
  geom_bar(aes(label, X2, fill = as.factor(groups)), 
           position = "dodge", stat = "summary", fun.y = "mean")

edited May 17 '16 at 05:50

PatrickT

10,037
9
76
111

answered May 12 '15 at 06:47

uhlitz

1,472
10
8

Wow this is awesome and really elegant. Thanks! Do you happen to know how to get darker colors in a nice way? – Cauchy May 12 '15 at 07:00
How do you modify this, to add labels on top of the bars? – Carrol Mar 30 '20 at 06:31
21

Tried this with `ggplot 3.3.2` and it seems like `fun.y` has been deprecated. Use `fun = "mean"` instead. – nniloc Jun 23 '20 at 17:52
1

Is this solution also compatible with geom_errorbar? – Rivered Jun 29 '22 at 11:37

score 6 · Answer 2 · edited May 17 '16 at 06:32

6

ggplot2 likes 1 data point for 1 plot point. Create a new data frame with your summary statistics, then plot with stat="identity"

require(reshape2)
plot.data <- melt(tapply(test2$X2, test2$groups,mean), varnames="group", value.name="mean")

 ggplot(plot.data, aes(x=group,y=mean)) + geom_bar(position="dodge", stat="identity")

edited May 17 '16 at 06:32

PatrickT

10,037
9
76
111

answered May 12 '15 at 06:28

Scransom

3,175
3
31
51

score 1 · Answer 3 · answered May 23 '18 at 11:00

Try using ggpubr. It creates ggplot2-like charts.

library(ggpubr)

ggbarplot(test2, x = "label", y = "X2",
          add = "mean", fill = "groups")

Alternatively, add a facet:

ggbarplot(test2, x = "label", y = "X2",
          add = "mean", fill = "groups",
          facet.by = "groups")

ggplot2 : Plot mean with geom_bar

3 Answers3

Linked

Related