22 Combining plots

We often need to combine plots to make a multipart figure in a manuscript or thesis.

First, check you actually need to combine plots and cannot use facets instead. Typically, use facets if the x-axis variable is the same for all plots and the plots use the same geoms.

Plots can be combined using the patchwork package.

library(patchwork)

Start by making some plots

p1 <- ggplot(penguins, aes(x = species, y = bill_length_mm)) +
  geom_boxplot()
p2 <- ggplot(penguins, aes(x = body_mass_g, y = bill_length_mm, colour = species)) +
  geom_point()
p3 <- ggplot(penguins, aes(x = body_mass_g, y = bill_length_mm, colour = species)) +
  geom_point() +
  facet_wrap(~ species)
p4 <- ggplot(penguins, aes(x = island, fill = species)) +
  geom_bar()

The simplest way to use patchwork is to + to combine the plots. patchwork will try to make the combined figure square.

p1 + p2 + p3 + p4

With | (side by side) \ (over under), you can have more control.

(p1 | p2  | p4) / p3

plot_layout gives you control over the relative size of the plot and whether the legends should be combined if possible.

(p1 | p2  | p4) / p3 + 
  plot_layout(heights = c(2, 1), guides = "collect")

If you want to change all the plots in the combined figure, you can add a ggplot2 function with &.

(p1 | p2  | p4) / p3 + 
  plot_layout(heights = c(2, 1), guides = "collect") &
  theme(panel.grid = element_blank())

You can also overlay figures, which might be useful to show an inset map.

library(rnaturalearth)
library(sf)

# map data inset
europe <- ne_countries(scale = 110, continent = "Europe", returnclass = "sf")

# map data main map
svalbard <- ne_countries(scale = 50, country = "Norway", returnclass = "sf") |>  
   st_crop(c(xmin = 0, xmax = 34, ymin = 76, ymax = 81))

# bounding box main map
bb <- svalbard |> 
  st_bbox()

#inset map
euro_map <- ggplot() +
  geom_sf(data = europe) +
  # annotate bounding box
  annotate(geom = "rect", 
           xmin = bb$xmin, xmax = bb$xmax, ymin = bb$ymin, ymax = bb$ymax, 
           fill = "red", alpha = 0.3) +
  coord_sf(xlim = c(-25, 40), y = c(50, 82)) +
  ggthemes::theme_map() +
  theme(
  #give white background and black border
    panel.background = element_rect(fill = "white", colour = "black"), 
    plot.margin = margin() #remove margins
  )

# main map
svalbard_map <- ggplot() +
  geom_sf(data = svalbard)

# combined map
svalbard_map + 
  inset_element(euro_map, 
                left = 0.7, 
                right = 0.99, 
                top = 0.4, 
                bottom = 0.01)

Contributors

Richard Telford

--- execute: warning: false --- ``` {r setup, include=FALSE} source("R/setup.R") ``` # Combining plots {#sec-combining-plots} We often need to combine plots to make a multipart figure in a manuscript or thesis. First, check you actually need to combine plots and cannot use facets instead. Typically, use facets if the x-axis variable is the same for all plots and the plots use the same geoms. Plots can be combined using the `patchwork` package. ```{r patchwork-load, message=FALSE} library(patchwork) ``` Start by making some plots ```{r some-plots} p1 <- ggplot(penguins, aes(x = species, y = bill_length_mm)) + geom_boxplot() p2 <- ggplot(penguins, aes(x = body_mass_g, y = bill_length_mm, colour = species)) + geom_point() p3 <- ggplot(penguins, aes(x = body_mass_g, y = bill_length_mm, colour = species)) + geom_point() + facet_wrap(~ species) p4 <- ggplot(penguins, aes(x = island, fill = species)) + geom_bar() ``` The simplest way to use `patchwork` is to `+` to combine the plots. `patchwork` will try to make the combined figure square. ```{r patchwork1} p1 + p2 + p3 + p4 ``` With `|` (side by side) `\` (over under), you can have more control. ```{r patchwork2} (p1 | p2 | p4) / p3 ``` `plot_layout` gives you control over the relative size of the plot and whether the legends should be combined if possible. ```{r patchwork-layout} (p1 | p2 | p4) / p3 + plot_layout(heights = c(2, 1), guides = "collect") ``` If you want to change all the plots in the combined figure, you can add a ggplot2 function with `&`. ```{r patchwork-ampersand} (p1 | p2 | p4) / p3 + plot_layout(heights = c(2, 1), guides = "collect") & theme(panel.grid = element_blank()) ``` You can also overlay figures, which might be useful to show an inset map. ```{r inset-map} library(rnaturalearth) library(sf) # map data inset europe <- ne_countries(scale = 110, continent = "Europe", returnclass = "sf") # map data main map svalbard <- ne_countries(scale = 50, country = "Norway", returnclass = "sf") |> st_crop(c(xmin = 0, xmax = 34, ymin = 76, ymax = 81)) # bounding box main map bb <- svalbard |> st_bbox() #inset map euro_map <- ggplot() + geom_sf(data = europe) + # annotate bounding box annotate(geom = "rect", xmin = bb$xmin, xmax = bb$xmax, ymin = bb$ymin, ymax = bb$ymax, fill = "red", alpha = 0.3) + coord_sf(xlim = c(-25, 40), y = c(50, 82)) + ggthemes::theme_map() + theme( #give white background and black border panel.background = element_rect(fill = "white", colour = "black"), plot.margin = margin() #remove margins ) # main map svalbard_map <- ggplot() + geom_sf(data = svalbard) # combined map svalbard_map + inset_element(euro_map, left = 0.7, right = 0.99, top = 0.4, bottom = 0.01) ``` ::: callout-note ## Further reading [patchwork package](https://patchwork.data-imaginist.com/) ::: ::: {.column-margin} ### Contributors {.unlisted .unnumbered} - Richard Telford :::