Colour, Fill, and Shape Palettes
In ggprism: A 'ggplot2' Extension Inspired by 'GraphPad Prism'

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(ggplot2)
library(ggprism)
library(patchwork)

ggprism includes colour and fill scales with various palettes that mirror the colour schemes available in GraphPad Prism.

Disclaimer: some of these palettes are quite nice, some are pretty ugly and not useful, and most are not colour blind friendly as far as I know (except palette = "colorblind_safe"). It would probably be safer to use other colour scales e.g. those in RColorBrewer for any serious application or work to be published.

In any case, these scales are used to colour the data elements of ggplots and can be used separately or together with theme_prism(). Additionally, a shape scale with 3 palettes is included which is similar to the shapes available in Prism. This vignette outlines how to use scale_colour_prism() (or scale_color_prism()), scale_fill_prism(), and scale_shape_prism().

Colour scales

First we'll create a base plot.

# create a base plot to compare colour scales
base <- ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point(aes(colour = factor(cyl), shape = factor(cyl)), size = 3) + 
  theme_prism() + 
  theme(legend.position = c(0.8, 0.8))

base

scale_colour_prism() (US spelling also works) functions pretty much identically to scale_colour_manual() from ggplot2. The default palette for scale_colour_prism() is "colors".

# compare manual colour scale with prism colour scale
p1 <- base + scale_colour_manual(values = c("blue", "red", "green3"))
p2 <- base + scale_colour_prism()

p1 + p2

There are several different colour palettes that are built into scale_colour_prism().

# see names and lengths of available scale_colour_prism() palettes
lengths(ggprism_data$colour_palettes)

Using the palettes is simple.

# try out some different colour palettes
p1 <- base + scale_colour_prism(palette = "purple_passion")
p2 <- base + scale_colour_prism(palette = "candy_bright")

p1 + p2

You can preview a single palette using the preview_theme() function. It shows a sample plot with the colour scale applied (as well as scale_fill_prism() and theme_prism()). See this page for images of most available colour palettes.

preview_theme("flames")

Fill scales

First we'll create a base plot.

# create a base plot to compare fill scales
base <- ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point(aes(fill = factor(cyl), shape = factor(cyl)), size = 3) + 
  theme_prism() + 
  theme(legend.position = c(0.8, 0.8)) + 
  scale_shape_prism(palette = "filled")

base

scale_fill_prism() functions pretty much identically to scale_fill_manual() from ggplot2. The default palette for scale_fill_prism() is "colors".

# compare manual fill scale with prism fill scale
p1 <- base + scale_fill_manual(values = c("blue", "red", "green3"))
p2 <- base + scale_fill_prism()

p1 + p2

There are several different fill palettes that are built into scale_fill_prism().

# see names and lengths of available scale_fill_prism() palettes
lengths(ggprism_data$fill_palettes)

Using the palettes is simple.

# try out some different fill palettes
p1 <- base + scale_fill_prism(palette = "colorblind_safe")
p2 <- base + scale_fill_prism(palette = "neon")

p1 + p2

You can preview a single palette using the preview_theme() function. It shows a sample plot with the colour scale applied (as well as scale_colour_prism() and theme_prism()). See this page for images of most available fill palettes.

preview_theme("diazo")

Shape scales

The order of shapes used in GraphPad Prism is slightly different to ggplot2, hence 3 different shape palettes are included in the ggprism function scale_shape_prism().

We can see the names and lengths of these shape palettes. Naturally, "default" is the default palette and does not include any shape with a fill. The "filled" palette should be used when you also want to use a fill scale, as the first 5 symbols have a fill. The "complete" palette symbols 5-9 have a fill.

# see names and lengths of available scale_shape_prism() palettes
lapply(ggprism_data$shape_palettes, nrow)

We can also define a function to see all the shapes in a specific palette.

# define a function for convenience
show_shapes <- function(palette) {
  df_shapes <- ggprism_data$shape_palettes[[palette]][, -1]
  df_shapes$pch_f <- factor(df_shapes$pch, levels = df_shapes$pch)

  ggplot(df_shapes, aes(x = 0, y = 0, shape = pch)) +
    geom_point(aes(shape = pch), size = 5, fill = 'red') +
    scale_shape_identity() +
    facet_wrap(~ pch_f) +
    theme_void()
}

# show the shapes in the palette "complete"
show_shapes("complete")

We'll define a base plot to see scale_shape_prism() in action.

# create a base plot to compare shape scales
base <- ggplot(mpg, aes(x = displ, y = cty)) +
  geom_point(aes(colour = class, fill = class, shape = class)) + 
  theme_prism(base_size = 11, base_fontface = "plain", border = TRUE) +
  theme(plot.subtitle = element_text(face = "bold"),
        legend.position = c(0.8, 0.75),
        legend.key.height = unit(10, "pt")) +
  coord_cartesian(clip = "off") + 
  scale_colour_prism(palette = "floral") + 
  scale_fill_prism(palette = "floral")

base

Then we'll compare the 3 shape palettes to the ggplot2 base plot.

# compare shape scales
p1 <- base
p2 <- base + scale_shape_prism(palette = "default") + 
  labs(subtitle = "default")
p3 <- base + scale_shape_prism(palette = "filled") + 
  labs(subtitle = "filled")
p4 <- base + scale_shape_prism(palette = "complete") + 
  labs(subtitle = "complete")

(p1 + p2) / (p3 + p4)