How to add a number of observations per group and use group mean in ggplot2 boxplot?

23
votes

I am doing a basic boxplot where y=age and x=Patient groups 

age <- ggplot(data, aes(factor(group2), age))  + ylim(15, 80) 
age + geom_boxplot(fill = "grey80", colour = "#3366FF")

I was hoping you could help me out with a few things:

1) Is it possible to include a number of observations per group above each group boxplot (but NOT on the X axis where my group labels are) without having to do this in paint :)? I have tried using:

age + annotate("text", x = "CON", y = 60, label = "25")

where CON is the 1st group and y = 60 is ~ just above the boxplot for this group. However, the command didn't work. I assume it has something to do that it reads x as a continuous rather than a categorical variable.

2) Also although there are plenty of questions about using the mean rather than the median for the boxplots, I still haven`t found a code that works for me?

3) On the same matter is there a way you could include the mean group stat in the boxplot? Perhaps using

age + stat_summary(fun.y=mean, colour="red", geom="point")

which however only includes a dot of where the mean lies. Or again using

age + annotate("text", x = "CON", y = 30, label = "30")

where CON is the 1st group and y = 30 is ~ the group age mean. Knowing how flexible and rich ggplot2 syntax is I was hoping that there is a more elegant way of using the real stats output rather than annotate.

Any suggestions/links would be much appreciated!

Thanks!!

3
rggplot2boxplot
A boxplot normally has min, lower, middle and upper quantiles and finally a max value. You already have the .25, .5 and .75 quantiles. Isn't this informative enough? – Arun
This is the format I am asked for. – user1442363

3 Answers

36
votes

Is this anything like what you're after? With stat_summary, as requested:

# function for number of observations 
give.n <- function(x){
  return(c(y = median(x)*1.05, label = length(x))) 
  # experiment with the multiplier to find the perfect position
}

# function for mean labels
mean.n <- function(x){
  return(c(y = median(x)*0.97, label = round(mean(x),2))) 
  # experiment with the multiplier to find the perfect position
}

# plot
ggplot(mtcars, aes(factor(cyl), mpg, label=rownames(mtcars))) +
  geom_boxplot(fill = "grey80", colour = "#3366FF") +
  stat_summary(fun.data = give.n, geom = "text", fun.y = median) +
  stat_summary(fun.data = mean.n, geom = "text", fun.y = mean, colour = "red")

Black number is number of observations, red number is mean value. joran's answer shows you how to put the numbers at the top of the boxes enter image description here

hat-tip: https://stackoverflow.com/a/3483657/1036500

18
votes

I think this is what you're looking for maybe?

myboxplot <- ddply(mtcars,
                    .(cyl),
                    summarise,
                    min = min(mpg),
                    q1 = quantile(mpg,0.25),
                    med = median(mpg),
                    q3 = quantile(mpg,0.75),
                    max= max(mpg),
                    lab = length(cyl))
ggplot(myboxplot, aes(x = factor(cyl))) + 
    geom_boxplot(aes(lower = q1, upper = q3, middle = med, ymin = min, ymax = max), stat = "identity") + 
    geom_text(aes(y = max,label = lab),vjust = 0)

enter image description here

I just realized I mistakenly used the median when you were asking about the mean, but you can obviously use whatever function for the middle aesthetic you please.

4
votes

Answer to the first problem. To show value above the box you should provide x values as numeric not as level names. So, to plot the value above first value give x=1.

data(ToothGrowth)
ggplot(ToothGrowth,aes(supp,len))+geom_boxplot()+
   annotate("text",x=1,y=32,label=30)