ggside R Package

The ggside package extends ggplot2 by allowing you to add side panels that display graphical information relative to the main panel's axes. This is useful for showing metadata for discrete axes or summary statistics (like boxplots or density distributions) for continuous axes.

Geometries follow a naming pattern: geom_xside* adds a layer to the x-side panel, and geom_yside* adds a layer to the y-side panel. By default, an xside geometry is positioned above the main panel and shares the same x-axis but has an independent y-axis.

library(ggplot2)
library(ggside)

ggplot(mpg, aes(displ, hwy, colour = class)) +
  geom_point(size = 2) +
  geom_xsidedensity(aes(y = after_stat(density)), position = "stack") +
  geom_ysidedensity(aes(x = after_stat(density)), position = "stack") +
  theme(axis.text.x = element_text(angle = 90, vjust = .5))

Since version 0.3.0, you can mix discrete and continuous axes without manually using scale_(y|x)side(x|y)_*() functions. ggside geometries automatically handle the mapping of positional aesthetics (x and y) to ysidex and xsidey under the hood.

ggplot(mpg, aes(displ, hwy, colour = class)) +
  geom_point(size = 2) +
  geom_xsideboxplot(aes(y =class), orientation = "y") +
  geom_ysidedensity(aes(x = after_stat(density)), position = "stack") +
  scale_ysidex_continuous(guide = guide_axis(angle = 90), minor_breaks = NULL) +
  theme(ggside.panel.scale = .3)

You can install the latest stable version of ggside from CRAN or install the development version directly from GitHub using devtools.

#CRAN
utils::install.packages("ggside")
#Github
devtools::install_github("jtlandis/ggside")

Adding side panels is similar to adding standard ggplot2 layers. Geometries follow a naming pattern: geom_xside* adds a geometry to the x-side panel, and geom_yside* adds a geometry to the y-side panel.

By default, an xside geometry (like geom_xsidedensity) is positioned above the main panel and shares the main panel's x-axis but maintains an independent y-axis.

library(ggplot2)
library(ggside)

ggplot(mpg, aes(displ, hwy, colour = class)) +
  geom_point(size = 2) +
  geom_xsidedensity(aes(y = after_stat(density)), position = "stack") +
  geom_ysidedensity(aes(x = after_stat(density)), position = "stack") +
  theme(axis.text.x = element_text(angle = 90, vjust = .5))

Since version 0.3.0, you can mix discrete and continuous axes without manually using scale_(y|x)side(x|y)_*() functions. ggside geometries automatically handle the necessary aesthetic mapping (ysidex and xsidey) under the hood. You can control the side panel size using theme(ggside.panel.scale = .3).

ggplot(mpg, aes(displ, hwy, colour = class)) +
  geom_point(size = 2) +
  geom_xsideboxplot(aes(y = class), orientation = "y") +
  geom_ysidedensity(aes(x = after_stat(density)), position = "stack") +
  scale_ysidex_continuous(guide = guide_axis(angle = 90), minor_breaks = NULL) +
  theme(ggside.panel.scale = .3)

A known issue occurs when a layer requires a positional scale on the main panel to be computed later, but the same scale is present on a parallel side layer meant to be discrete. This can result in warnings and missing data.

Problematic pattern:

ggplot(iris, aes(Sepal.Width)) +
  geom_density() +
  geom_xsidepoint(aes(y = Species))

Solution: Explicitly declare the scale for the main panels to ensure proper initialization.

ggplot(iris, aes(Sepal.Width)) +
  geom_density() +
  geom_xsidepoint(aes(y = Species)) +
  scale_y_continuous()