Lecture 5.2 KEY

Note

This template follows lecture 5.2 slides. Please be sure to cross-reference the slides, which contain important information and additional context!

Setup

##~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
##                                    setup                                 ----
##~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

#..........................load packages.........................
library(palmerpenguins)
library(tidyverse)

##~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
##                              create base plots                           ----
##~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

# requires categorical color scale ----
cat_color_plot <- ggplot(na.omit(penguins), 
                         aes(x = bill_length_mm, y = bill_depth_mm, 
                             color = species, shape = species)) +
  geom_point(size = 4, alpha = 0.8)

# requires continuous color scale ----
cont_color_plot <- ggplot(na.omit(penguins), 
                          aes(x = bill_length_mm, y = bill_depth_mm, 
                              color = body_mass_g)) +
  geom_point(size = 4, alpha = 0.8) 

Colors for inclusive & accessible design

Viridis scales

• Check out the documentation for additional viridis palette options.

# discrete viridis scales ----
cat_color_plot +
  scale_color_viridis_d(option = "viridis") 

# continuous color scales ----
cont_color_plot +
  scale_color_viridis_c(option = "magma")

RColorBrewer scales

• Exploring available palettes and generating HEX codes (or check out the web-based interface):

# display only colorblind-friendly RColorBrewer palettes ----
RColorBrewer::display.brewer.all(colorblindFriendly = TRUE)

# preview palette with your number of desired colors ----
RColorBrewer::display.brewer.pal(n = 3, name = 'Dark2')

# print the HEX codes of your palette ----
RColorBrewer::brewer.pal(n = 3, name = 'Dark2')

[1] "#1B9E77" "#D95F02" "#7570B3"

# or save hex codes as a vector ----
my_brewer_pal <- RColorBrewer::brewer.pal(n = 3, name = 'Dark2')

• Using RColorBrewer palettes

# for qualitative palettes ----
cat_color_plot +
  scale_color_brewer(palette = "Dark2") 

# for unclassed continuous color scales ----
cont_color_plot +
  scale_color_distiller(palette = "BuPu")

# for classed continuous color scales ----
cont_color_plot +
  scale_color_fermenter(palette = "YlGnBu")

Accessibility tips

• outline light-colored points, which are difficult to see

ggplot(penguins, aes(x = bill_length_mm, y = bill_depth_mm, fill = body_mass_g)) +
  geom_point(shape = 21, size = 4, alpha = 0.8) +
  scale_fill_distiller(palette = "BuPu")

• use redundant mapping (colors & shapes) whenever possible (recall our original plot, with an added viridis color palette)

ggplot(penguins, aes(x = bill_length_mm, y = bill_depth_mm, color = species, shape = species)) +
  geom_point(size = 4, alpha = 0.8) + 
  scale_color_viridis_d(option = "turbo") 

Paletteer

I. apply palette using scale_*_paletteer_*()

# discrete data / palette ----
cat_color_plot +
  paletteer::scale_color_paletteer_d("calecopal::superbloom3")

# continuous data / palette (unclassed) ----
cont_color_plot +
  paletteer::scale_color_paletteer_c("scico::batlow", direction = -1)

# continuous data / palette (classed) ----
cont_color_plot +
  paletteer::scale_color_paletteer_binned("scico::batlow", direction = -1)

II. create vector of colors using paletteer_*(), then apply using the appropriate ggplot::scale_*() function

• discrete data / palette example:

# create palette ----
pal_d <- paletteer::paletteer_d("wesanderson::GrandBudapest1", n = 3)
pal_d 

<colors>
#F1BB7BFF #FD6467FF #5B1A18FF 

# apply to scatter plot (use `color` variant) ----
cat_color_plot + 
  scale_color_manual(values = pal_d)

# apply to histogram (use `fill` variant) ----
ggplot(penguins, aes(x = body_mass_g, fill = species)) +
  geom_histogram() +
  scale_fill_manual(values = pal_d)

• continuous data / palette example:

# create palette ----
pal_c <- paletteer::paletteer_c("ggthemes::Gold-Purple Diverging", n = 5)
pal_c

<colors>
#AD9024FF #CBAE4DFF #E3D8CFFF #CA96B9FF #AC7299FF 

# apply to scatter plot as an unclassed palette (use `gradientn` variant) ---- 
cont_color_plot + 
  scale_color_gradientn(colors = pal_c)

# apply to scatter plot as a classed (binned) palette (use `stepsn` variant) ---- 
cont_color_plot + 
  scale_color_stepsn(colors = pal_c)

Subduing pure hues

1. Adjust saturation when selecting a HEX code from Google’s color picker, https://g.co/kgs/9SQkdgv, or adjust chroma directly in your ggplot:

ggplot(na.omit(penguins), aes(x = species, fill = sex)) +
  geom_bar() +
  scale_fill_hue(c = 70)

2. Adjust value when selecting a HEX code from Google’s color picker, https://g.co/kgs/9SQkdgv, or adjust lightness directly in your ggplot:

ggplot(na.omit(penguins), aes(x = bill_length_mm, y = bill_depth_mm, color = sex)) +
  geom_point() +
  scale_color_hue(l = 45)

3. Increase transparency using the alpha argument in ggplot

ggplot(penguins, aes(x = species)) +
  geom_bar(fill = "#00FF33", color = "gray7", alpha = 0.5) + 
  theme_classic() 

Some final palette tips

Save palette outside of plot

# create palette ----
my_palette <- c("#32DE8A", "#E36414", "#0F4C5C")

# apply to plot ----
cat_color_plot +
  scale_color_manual(values = my_palette) 

Set color names

# create palette ----
my_palette_named <- c("Adelie" = "#32DE8A","Chinstrap" = "#E36414", "Gentoo" = "#0F4C5C")

# apply to plot (all penguins) ----
ggplot(penguins, aes(x = bill_length_mm, y = bill_depth_mm, color = species)) +
  geom_point(size = 4, alpha = 0.8) +
  scale_color_manual(values = my_palette_named)

# apply to plot (just adelie & gentoo) ----
penguins |> 
  filter(species != "Chinstrap") |> 
  ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = species)) +
  geom_point(size = 4, alpha = 0.8) +
  scale_color_manual(values = my_palette_named)

Use scale_*_identity()

# example 1 (color points based on value) ----
penguins |> 
  mutate(
    my_color = case_when(
      bill_length_mm < 40 ~ "#D7263D",
      between(bill_length_mm, 40, 50) ~ "#E4BB97",
      bill_length_mm > 50 ~ "#386150"
    )
  ) |> 
  ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = my_color)) +
  geom_point(size = 4, alpha = 0.8) +
  scale_color_identity()

# example 2 (highlight points based on criteria)
penguins |> 
  mutate(
    my_color = case_when(
      body_mass_g > 6000 ~ "#D7263D",
      TRUE ~ "gray50"
    )
  ) |> 
  ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = my_color)) +
  geom_point(size = 4, alpha = 0.8) +
  scale_color_identity(guide = "legend", 
                       name = "Body mass (g)", labels = c(">6000", "<= 6000"))