Packages

For this post, we need to install and load the beeswarm package.

We can install it from CRAN using install.packages("beeswarm"). Then, we can load it:

# install.packages("beeswarm")
library(beeswarm)

Dataset

Since beeswarm plots are made to visualize individual data points, we need a dataset that contains numerical values. Here, we’ll use the iris dataset, which is a built-in dataset in R.

We can easily load it:

data(iris)

Default grouped beeswarm plot

The package comes with a beeswarm() function, and thanks to the ~ operator, we can easily create a grouped beeswarm plot. Here, we’ll use the Species column to group the data:

The following code basically means:

plot Sepal.Length for each Species using the iris dataset.

beeswarm(
  Sepal.Length ~ Species,
  data=iris
)

Use different colors

Even tough the default option does what we want, having the same color for all the Species makes the plot less readable. We can use the col argument to change the color of the dots:

beeswarm(
  Sepal.Length ~ Species, 
  data=iris,
  col=c("orange", "lightblue", "magenta"),
  pch = 19 # fill the dots
)

Custom position behavior

When you have lots of data points, it can be useful to change the position behavior of the dots in order to avoid overlapping between groups.

Fortunately, the corral argument allows you to change this. The available options are:

• none (default): no correction
• gutter: fix a higher and lower limit for each group
• wrap: similar to gutter but add random noise to the position of the dots
• random: randomly position the dots
• omit: omit the dots that would overlap

Here’s an example with the gutter option:

beeswarm(
  Sepal.Length ~ Species, 
  data=iris,
  col=c("orange", "lightblue", "magenta"),
  pch = 19, # fill the dots
  corral = "gutter"
)

Going further

This post explains how to create a grouped beeswarm plot with R.

You might also be interested in how to create a beeswarm plot with ggplot2.

Related chart types

---

Violin

Density

Histogram

Boxplot

Ridgeline