What if my data has more categories than the number of colors offered?

Although it is recommended that we do not encode too many categories in different colors, in practice, one can still create an “adaptive” color palette based on the existing discrete color palettes in ggsci.

See this blog post for a detailed guide on creating adaptive ggplot2 color scales with color interpolation.

Use a color scale consistently for multiple plots in a document

To apply a color scale for all plots in a document and avoid repetition, a simple solution is setting the two global options ggplot2.discrete.colour and ggplot2.discrete.fill. For example:

library("ggplot2")

p <- ggplot(mpg, aes(displ, hwy, colour = factor(cyl), fill = factor(cyl))) +
  geom_point() +
  geom_smooth(method = "lm") +
  theme_bw()

p

# Set global options
options(
  ggplot2.discrete.colour = ggsci::scale_colour_d3,
  ggplot2.discrete.fill = ggsci::scale_fill_d3
)

p

# Restore original options after use
options(
  ggplot2.discrete.colour = NULL,
  ggplot2.discrete.fill = NULL
)

p