Split violin plots

June 25, 2013
By

(This article was first published on Ecology in silico, and kindly contributed to R-bloggers)

Violin plots are useful for comparing distributions. When data are
grouped by a factor with two levels (e.g. males and females), you can
split the violins in half to see the difference between groups. Consider
a 2 x 2 factorial experiment: treatments A and B are crossed with groups
1 and 2, with N=1000.


1
2
3
4
5
6
7
8
9
# Simulate data
n.each <- 1000
A1 <- rnorm(n.each, 2, 1)
A2 <- rnorm(n.each, 1.5, 2)
B1 <- rnorm(n.each, 4, 1.5)
B2 <- rnorm(n.each, 0, 1)
values <- c(A1, A2, B1, B2)
treatment <- rep(c("A", "B"), each=n.each*2)
group <- rep(c(1, 2, 1, 2), each=n.each)

Boxplots are often used:


1
2
par(bty="n")
boxplot(values ~ group*treatment, main="Box plot", col=rep(c("purple", "lightblue"), 2))

This gives us a rough comparison of the distribution in each group,
but sometimes it’s nice to visualize the kernel density estimates instead.

I recently ran into this issue and tweaked the vioplot() function from
the vioplot
package by Daniel Adler to make split violin plots.
With vioplot2(), the side
argument specifies whether to plot the density on “both”, the “left”, or
the “right” side.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
require(vioplot)
require(devtools)
require(digest)
source_gist("https://gist.github.com/mbjoseph/5852613")
plot(x=NULL, y=NULL,
     xlim = c(0.5, 2.5), ylim=c(min(values), max(values)),
     type="n", ann=FALSE, axes=F)
axis(1, at=c(1, 2),  labels=c("A", "B"))
axis(2)
for (i in unique(treatment)) {
  for (j in unique(group)){
    vioplot2(values[which(treatment == i & group == j)],
             at = ifelse(i == "A", 1, 2),
             side = ifelse(j == 1, "left", "right"),
             col = ifelse(j == 1, "purple", "lightblue"),
             add = T)
  }
}
title("Violin plot", xlab="Treatment")
legend("bottomright", fill = c("purple", "lightblue"),
       legend = c("Group 1", "Group 2"), box.lty=0)

Last but not least, Peter Kampstra’s
beanplot
package uses beanplot() to make split
density plots, but 1) plots a rug rather
than a quantile box, 2) includes a line for the overall mean or median,
and 3) makes it easier to change the kernel function.


1
2
3
4
5
6
7
8
9
require(beanplot)
beanplot(values ~ group*treatment, ll = 0.04,
         main = "Bean plot", side = "both", xlab="Treatment",
         col = list("purple", c("lightblue", "black")),
         axes=F)
axis(1, at=c(1, 2),  labels=c("A", "B"))
axis(2)
legend("bottomright", fill = c("purple", "lightblue"),
       legend = c("Group 1", "Group 2"), box.lty=0)

There are
more
ways
than
one
to
skin
a
cat,
and what one uses will probably come to personal preference.

To leave a comment for the author, please follow the link and comment on their blog: Ecology in silico.

R-bloggers.com offers daily e-mail updates about R news and tutorials on topics such as: Data science, Big Data, R jobs, visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...



If you got this far, why not subscribe for updates from the site? Choose your flavor: e-mail, twitter, RSS, or facebook...

Comments are closed.

Sponsors

Mango solutions



plotly webpage

dominolab webpage



Zero Inflated Models and Generalized Linear Mixed Models with R

Quantide: statistical consulting and training

datasociety

http://www.eoda.de





ODSC

ODSC

CRC R books series





Six Sigma Online Training









Contact us if you wish to help support R-bloggers, and place your banner here.

Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)