ggspruce

ggspruce provides functions that make automatic adjustments to ggplot2 plots to improve aesthetics.

Installation

You can install the development version of ggspruce from GitHub with:

# install.packages("devtools")
devtools::install_github("sheridar/ggspruce")

Optimizing color palettes

This is a color palette from the MetBrewer package. This palette can be improved by modifying colors 2, 12, and 13 to make them more distinct.

library(ggspruce)
library(ggplot2)
library(MetBrewer)
library(cowplot)

clrs <- as.character(met.brewer("Austria", 15))

plot_colors(clrs)

The spruce_colors() function calculates the pairwise distance between colors in the palette, and identifies pairs of colors that are very similar. The difference threshold used to identify similar colors can be modified with the difference parameter, with higher values resulting in colors that are more distinct. How colors are modified can be controlled using the property parameter.

By default, new colors are interpolated based on the other colors on the palette.

new_clrs <- clrs |>
  spruce_colors(difference = 10)

plot_colors(new_clrs)

One or more color properties can also be specified to limit how the colors are adjusted. If the example below lightness is first adjusted, if the resulting colors still do not meet the difference threshold, the hue is adjusted.

new_clrs <- clrs |>
  spruce_colors(
    difference = 20,
    property   = c("lightness", "hue")
  )

plot_colors(new_clrs)

Integration with ggplot2

Color palettes can be automatically modified on the fly using the scale_color_spruce() and scale_fill_spruce() functions. These functions are similar to the ggplot2::scale_*_manual() functions with some modifications. Original colors are shown on the left and adjusted colors on the right.

clrs <- as.character(met.brewer("Nattier", 8))

# Original colors
plt <- diamonds |>
  ggplot(aes(carat, price, color = clarity)) +
  geom_point(size = 1) +
  guides(color = guide_legend(override.aes = list(size = 3))) +
  theme_bw() +
  theme(
    aspect.ratio = 1,
    legend.position = "bottom"
  )

plt1 <- plt +
  scale_color_manual(values = clrs) +
  ggtitle("original")

# Automatically adjust colors
plt2 <- plt +
  scale_color_spruce(
    values     = clrs,
    difference = 25,
    property   = "lightness"
  ) +
  ggtitle("adjusted")

plot_grid(plt1, plt2, nrow = 1)
#> ℹ The minimum color difference for the adjusted palette is 23, increase `maxit` to improve optimization.

If no colors are provided to the scale_*_spruce() functions, they will pull colors from the existing color scale.

# Adjust colors from existing scale
plt1 <- plt +
  scale_color_discrete()

plt2 <- plt1 +
  scale_color_spruce(difference = 25)

plot_grid(plt1, plt2, nrow = 1)

Colorblind filters

Color similarity can be assessed and adjusted using colorblind filters. The original and adjusted palettes are shown below.

library(colorspace)
library(RColorBrewer)

clrs <- RColorBrewer::brewer.pal(10, "RdYlGn")

new_clrs <- clrs |>
  spruce_colors(
    difference = 15,
    property   = c("lightness", "saturation"),
    filter     = "deutan"
  )
#> ℹ The minimum color difference for the adjusted palette is 12.9, increase `maxit` to improve optimization.

# Plot adjusted colors without filter
plt1 <- clrs |>
  plot_colors() +
  ggtitle("original")

plt2 <- new_clrs |>
  plot_colors() +
  ggtitle("adjusted")

plot_grid(plt1, plt2, ncol = 1)

The palettes are shown below using a filter to simulate deuteranopia.

# Plot adjusted colors with colorblind filter
plt3 <- clrs |>
  deutan() |>
  plot_colors() +
  ggtitle("original + filter")
  
plt4 <- new_clrs |>
  deutan() |>
  plot_colors() +
  ggtitle("adjusted + filter")
  
plot_grid(plt3, plt4, ncol = 1)