7.7.1 Facet wrapping

Facet wrapping separates subsets of a dataset into plots with identical axes. This can be useful if different groups of data points overlap and occlude one another. Consider the following dataset with three groups of normally distributed random samples:

library(mvtnorm) # package implementing multivariate normal distributions
nsamp <- 100
data <-  rbind(
    rmvnorm(nsamp,c(1,1),sigma=matrix(c(1,0.8,0.8,1),nrow=2)),
    rmvnorm(nsamp,c(1,1),sigma=matrix(c(1,-0.8,-0.8,1),nrow=2)),
    rmvnorm(nsamp,c(1,1),sigma=matrix(c(1,0,0,1),nrow=2))
)
colnames(data) <- c('x','y')
g_oneplot <- as_tibble(data) %>%
  mutate(
    sample_name=c(rep('A',nsamp),rep('B',nsamp),rep('C',nsamp))
  ) %>%
  ggplot(aes(x=x,y=y,color=sample_name)) +
  geom_point()
g_oneplot

The three sample data points are plotted on top of each other, and although they are colored differently, it is difficult to discern the patterns of each group clearly. This same plot may be split into three, one for each sample group, using the facet_wrap() function:

g_oneplot + facet_wrap(vars(sample_name))

Now the trends of each group are clearly visible. The facet_wrap() function accepts one required argument, which is the variable name in the data that should be used to separate the groups decorated by the vars() function. Any geometry may be used with a facet wrap. For example, we could add a geom_smooth() to the plot to estimate a smoothing function within each facet:

g_oneplot + facet_wrap(vars(sample_name)) +
  geom_smooth(method="loess", formula=y ~ x)

By default, facet wrapping fixes all the axes of each plot to be identical, which enables the most accurate estimates of comparisons between plots according to the Elementary Perceptual Tasks hierarchy. This and many other aspects of the faceting may be adjusted using arguments as described in the documentation.

7.7.2 Multipanel Figures

Facet wrapping is useful when all of the data for the separate plots are on the same scale, and comaprisons of the facet groups is meaningful. However, faceting is less convenient when we desire to place plots with unrelated data in the same figure. The patchwork package allows us to compose multiple plots together using an intuitive set of operators. For example, we may put two plots next to each other by separating them with the | operator:

data <- tibble(
  a=rnorm(100,0,1),
  b=rnorm(100,3,2)
)
g_scatter <- ggplot(data, aes(x=a, y=b)) +
  geom_point()
g_violin <- pivot_longer(data, c(a,b)) %>% ggplot(aes(x=name,y=value,fill=name)) +
  geom_violin()

g_scatter | g_violin

Note that each plot was saved to its own variable and then composed with the patchwork operator |, which puts the plots side-by-side. We could also place one plot on top of the other with the / operator:

g_scatter / g_violin

Plots may be composed into arbitrarily complex multipanel figures in this way using () to group plots together:

g_scatter / ( g_scatter | g_violin)

Another possible composition:

(g_scatter / g_scatter ) | g_violin

The patchwork library is very flexible, and allows the designer to compose plots in many different ways. These multipanel plots make excellent starting points for producing publication-quality figures, as described in the next section.