patchwork in R [layouts, shared legends and inset plots]

Package

Author

Thomas Lin Pedersen

Sample plots

The following four plots will be used throughout this tutorial. Save them as objects so patchwork can combine them. If you want to compare patchwork with other alternatives such as cowplot or gridExtra, check the guide about combining plots in ggplot2.

# install.packages("ggplot2")
library(ggplot2)

set.seed(5)
x <- runif(100)
df <- data.frame(x = seq_along(x), y = x)

p1 <- ggplot(df, aes(x = "", y = y)) +
  geom_boxplot()

p2 <- ggplot(df, aes(y = y)) +
  geom_histogram(bins = 15)

p3 <- ggplot(df, aes(x = x, y = y)) +
  geom_line()

p4 <- ggplot(df, aes(x = x, y = y)) +
  geom_point()

Four ggplot2 sample plots

Operators

Load patchwork and use the + operator to place plots side by side. With two plots the result is identical to |, but with more plots + tries to create a square layout.

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

p1 + p2

Combine two ggplot2 plots side by side with patchwork

Place ggplot2 plots in a single row with the pipe operator in patchwork

| forces a single row, regardless of the number of plots.

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

p1 | p2 | p3

/ stacks plots in a single column.

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

p1 / p2

Stack ggplot2 plots with the slash operator in patchwork

Nesting operators

Operators can be nested with parentheses to create complex layouts without any extra function calls.

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

(p1 | p2) / p3

Nest patchwork operators to create complex ggplot2 layouts

One plot left two plots right with patchwork

One plot at the left, two stacked at the right.

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

p1 | (p2 / p3)

`plot_layout()`

plot_layout() controls the number of rows and columns, and the relative sizes of each plot with widths and heights.

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

p1 + p2 + p3 + p4 +
  plot_layout(ncol = 3, widths = c(1, 1, 2))

Control number of columns and relative widths with plot_layout in patchwork

Custom design

Pass a string to design where each character is a plot number and # is an empty area. This gives you full control over the layout.

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

design <- "
  111
  2#3
  443
"

p1 + p2 + p3 + p4 +
  plot_layout(design = design)

Custom layout design string in patchwork

Shared legend

When plots share the same legend, pass guides = "collect" to plot_layout() to merge them into a single legend for the whole figure.

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

set.seed(1)
df2 <- data.frame(x = 1:50, y = rnorm(50),
                  g = rep(c("A", "B"), 25))

pa <- ggplot(df2, aes(x, y, color = g)) + geom_point()
pb <- ggplot(df2, aes(x, y, color = g)) + geom_line()

pa + pb +
  plot_layout(guides = "collect")

Shared legend for combined ggplot2 plots with patchwork guides collect

Move shared legend to the bottom of combined ggplot2 plots

Changing the legend position

Use the & operator to apply theme() to all plots at the same time. This lets you move the collected legend to any side, as with the legend position of a single plot.

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

set.seed(1)
df2 <- data.frame(x = 1:50, y = rnorm(50),
                  g = rep(c("A", "B"), 25))

pa <- ggplot(df2, aes(x, y, color = g)) + geom_point()
pb <- ggplot(df2, aes(x, y, color = g)) + geom_line()

pa + pb +
  plot_layout(guides = "collect") &
  theme(legend.position = "bottom")

`plot_annotation()`

Add a title, subtitle and caption for the whole figure with plot_annotation().

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

p1 + p2 +
  plot_annotation(
    title = "Figure title",
    subtitle = "Subtitle",
    caption = "Caption")

Add a title to combined ggplot2 plots with plot_annotation in patchwork

Label each plot automatically with tag_levels in patchwork

Automatic plot labels

tag_levels adds a label to each panel automatically. Options are "1", "a", "A", "i" or "I" for Roman numerals.

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

p1 + p2 + p3 +
  plot_annotation(tag_levels = "A")

`&` operator

The & operator applies a layer to all plots at once. Use it to set the same theme, color scale or any other layer across the whole figure.

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

p1 + p2 + p3 + p4 & theme_minimal()

Apply the same theme to all ggplot2 plots at once with the & operator in patchwork

`inset_element()`

inset_element() places a plot on top of another at any position. The left, bottom, right and top arguments set the location as fractions of the parent plot area (0 to 1).

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

p3 + inset_element(p1,
                   left = 0.6,
                   bottom = 0.6,
                   right = 1,
                   top = 1)

Inset a ggplot2 plot inside another with inset_element from patchwork

Inset plot aligned to the panel area with align_to in patchwork

align_to

By default coordinates are relative to the full plot area including margins. Set align_to = "panel" to position the inset relative to the inner panel area instead.

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

p3 + inset_element(p1,
                   left = 0,
                   bottom = 0.55,
                   right = 0.4,
                   top = 1,
                   align_to = "panel")