In ab604/sotoncolours: University of Southampton colour palettes

knitr::opts_chunk$set(fig.width = 7, 
                      echo = FALSE, 
                      warning = FALSE, 
                      message=FALSE)

library(tidyverse)
library(sotoncolours)
library(dichromat)
library(colorspace)
library(palmerpenguins)
library(raster)
library(patchwork)

Introduction

The sotoncolours package contains colour palettes based on the University of Southampton colour scheme.

Installation

devtools::install_github("ab604/sotoncolours", build_vignettes = TRUE)

Colour and palette choices

The package contains 19 colours and 7 colour palettes contained within a named character vector called uos_colours:

<<<<<<< HEAD

# Named character vector of the colours
uos_colours

# See names of all palettes
names(uos_palettes)

To see the colours in a palette by name

# See names of all palettes
names(uos_palettes$contrast1)

pal_names <- names(uos_palettes)
par(mfrow=c(length(uos_palettes)/2, 2), lheight = 2, mar=rep(1, 4), adj = 0)
for (i in 1:length(uos_palettes)){
    viz_palette(uos_palettes[[i]], pal_names[i])
}

Colour vision impairment

These plots show how each of the palettes perform for different types of colour vision impairment (using dichromat).

Green-Blind (Deuteranopia)

par(mfrow=c(length(uos_palettes)/2, 2), lheight = 2, mar=rep(1, 4), adj = 0)
for (i in 1:length(uos_palettes)){
    viz_palette(dichromat(uos_palettes[[i]], "deutan"), pal_names[i])
}

Red-Blind (Protanopia)

par(mfrow=c(length(uos_palettes)/2, 2), lheight = 2, mar=rep(1, 4), adj = 0)
for (i in 1:length(uos_palettes)){
    viz_palette(dichromat(uos_palettes[[i]], "protan"), pal_names[i])
}

Blue-Blind (Tritanopia)

par(mfrow=c(length(uos_palettes)/2, 2), lheight = 2, mar=rep(1, 4), adj = 0)
for (i in 1:length(uos_palettes)){
    viz_palette(dichromat(uos_palettes[[i]], "tritan"), pal_names[i])
}

Desaturated

par(mfrow=c(length(uos_palettes)/2, 2), lheight = 2, mar=rep(1, 4), adj = 0)
for (i in 1:length(uos_palettes)){
    viz_palette(desaturate(uos_palettes[[i]]), pal_names[i])
}

Usage

Colours can be called directly using uos_cols() with a character argument for the colour name e.g. "horizon5".

ggplot(mtcars, aes(hp, mpg)) +
    geom_point(color = uos_cols("horizon5"),
               size = 4, alpha = .8)

The package contains color scale functions for ggplot2 plots: scale_color_uos() and scale_fill_uos() calling one of the palettes.

Here the marine and horizon5 palettes are used for a scatter plot of mtcars data.

# Use the marine palette for colours
p1 <- ggplot(mtcars, aes(mpg, wt)) + 
   geom_point(aes(colour = factor(cyl)), size=2, alpha=0.7) + 
   scale_colour_uos(palette="marine") +
    theme_bw() + 
    theme(aspect.ratio=1, legend.position="bottom")

# Use the horizon palette for colours
p2 <- ggplot(data = mpg) + 
   geom_point(mapping = aes(x = displ, y = hwy, color = class), alpha=0.7) +
   scale_colour_uos(palette="horizon") +
    theme_bw() + 
    theme(aspect.ratio=1, legend.position="bottom")

p1 + p2

Here the default palette marine is used in a stacked bar chart using the diamonds data.

 ggplot(diamonds) + 
  geom_bar(aes(x = cut, fill = clarity)) +
  scale_fill_uos() +  # defaults to marine
  theme_bw()

It's possible, but messy, to pick any colour combination from uos_colours, but the syntax requires extracting the hex codes using unname() and subsetting uos_cols using a character vector of the desired colour names.

It may be easier to create an object for any bespoke palette and pass then pass it to the respective ggplot function.

Here's an example with the Palmer penguin data using unname(uos_colours[c("black","shamrock","horizon4")]) as described above by creating a new palette called my_cols that is passed to the values argument of scale_colour_manual().

# Create a new vector of colours to pass to scale_colour manual
# Use unname() to get hex codes from uos_cols()
my_cols <- unname(uos_colours[c("black","shamrock","horizon4")])

ggplot(data = penguins, aes(x = species, y = flipper_length_mm)) +
  geom_boxplot(aes(color = species), width = 0.3, show.legend = FALSE) +
  geom_jitter(aes(color = species), alpha = 0.7, show.legend = FALSE, 
              position = position_jitter(width = 0.2, seed = 0)) +
  scale_colour_manual(values = my_cols) + # pass my_cols to scale_colour_manual
  theme_minimal() +
  labs(x = "Species",
       y = "Flipper length (mm)")

Here is a base R plot using the horizon palette directly via the colorRampPalette() function.

pal <- colorRampPalette(uos_palettes[["horizon"]])
image(volcano, col = pal(20))

Here is the same plot, but by creating a new palette from specific colours as with the penguin data.

# Create new palette
my_cols <- unname(uos_cols()[c('marine1','shamrock','horizon3')])

pal <- colorRampPalette(my_cols)
image(volcano, col = pal(20))