18 Scales, Formatting & Color Accessibility

In this chapter, we will format axis labels with the scales package, use colorblind-safe viridis palettes and check contrast for accessibility.

Why this matters

Raw numbers like 150000 slow readers down. Formatted labels ($150k, 45%) and palettes readable by colorblind viewers make charts instantly legible to everyone.

Common pitfall

Red-green contrasts are invisible to the most common form of colorblindness. Never encode critical groups with a red-versus-green scale alone.

18.1 Percent Labels

label_percent() turns 0–1 fractions into readable percentages:

df <- data.frame(group = c("A", "B", "C"), share = c(0.45, 0.30, 0.25))
ggplot(df) +
  geom_bar(aes(group, share), stat = "identity", fill = "steelblue") +
  scale_y_continuous(labels = label_percent())

18.2 Dollar Labels

label_dollar() formats currency axes:

sales <- data.frame(month = 1:6, revenue = c(12000, 15000, 13500, 18000, 21000, 19500))
ggplot(sales) +
  geom_line(aes(month, revenue), linewidth = 1) +
  geom_point(aes(month, revenue), size = 2) +
  scale_y_continuous(labels = label_dollar())

18.3 Comma Labels

label_comma() keeps large counts readable:

counts <- data.frame(year = 2019:2024, users = c(120000, 150000, 148000, 190000, 230000, 265000))
ggplot(counts) +
  geom_bar(aes(factor(year), users), stat = "identity", fill = "steelblue") +
  scale_y_continuous(labels = label_comma())

18.4 Viridis Palettes

Viridis palettes are perceptually uniform and readable by most colorblind viewers. Use them for continuous fills:

ggplot(mtcars) +
  geom_point(aes(disp, mpg, color = hp)) +
  scale_color_viridis_c()

And for discrete groups:

ggplot(mtcars) +
  geom_point(aes(disp, mpg, color = factor(cyl))) +
  scale_color_viridis_d()

18.5 Contrast Check

Before publishing, verify your figure in grayscale: print it or desaturate it. If groups are still distinguishable by luminance, the contrast passes. The scale_*_viridis_*() palettes pass this test by design; hand-picked pastels often fail it.

Where to go next

scales package reference: formatting functions gallery
Claus Wilke, Fundamentals of Data Visualization: color chapters
Colorblind simulator (e.g. Coblis) to check exported figures