Boxplot in R pattern

Question

I intend to connect a line connecting several boxplots on R. I’ve tried using the functions abline and lines, but I did not succeed. How can I solve my problem?

Answer 1

As no data was provided to solve the problem, I will use the data set iris to solve it. This data set has 150 observations on 4 quantitative variables and 1 categorical. This categorical variable has three levels. Because of these characteristics, this data set becomes very interesting to illustrate how to solve this problem.

Also, I will solve it in two ways: with the function boxplot, which is standard of R, and with the package ggplot2.

Boxplot Standard

First we need to plot the simple boxplot, using a quantitative variable as response and a categorical variable as predictor:

boxplot(Petal.Length ~ Species, data=iris)

Next, it is necessary to calculate the median for each plant species, so that the boxplots are connected according to some criterion. One way to do this is with the Aggregate function:

aggregate(iris$Petal.Length, list(iris$Species), median)
     Group.1    x
1     setosa 1.50
2 versicolor 4.35
3  virginica 5.55

For the graph to come out more easily, I will save this result inside an object called medianas. From that, just use lines and plot the averages calculated on top of the generated boxplot, considering that each variable level Species can be understood as a number from 1 to 3:

medianas <- aggregate(iris$Petal.Length, list(iris$Species), median)
lines(1:3, medianas$x)

Boxplot with ggplot2

I find this solution more elegant because the ggplot2 calculates and plots alone the statistics we want to put on the chart. To do this, just use the function stat_summary:

library(ggplot2)

ggplot(iris, aes(x=Species, y=Petal.Length)) +
  geom_boxplot() +
  stat_summary(fun.y=median, geom="line", lwd=1, aes(group=1))