ggplot2: Change color of axis text to match group

December 6, 2023

Using tidyverse and the mtcars dataset, I am plotting the model and mpg of each car. I would like the axis text color of each car model to reflect the make of the car (for example, Toyota = red, Cadillac = blue…etc). How can I do this?

Example Dataset

library(tidyverse)

# Create make and model for each car
mtcars <- mtcars %>%
  rownames_to_column(var = 'car') %>%
  filter(!car %in% c('Mazda RX4 Wag', 'Hornet 4 Drive', 'Valiant', 'Ford Pantera L')) %>%
  separate_wider_delim(car, " ", names = c('make', 'model'), cols_remove = FALSE) %>%
  select(car, make, model, mpg)

head(mtcars)
#> # A tibble: 6 × 4
#>   car               make   model        mpg
#>   <chr>             <chr>  <chr>      <dbl>
#> 1 Mazda RX4         Mazda  RX4         21  
#> 2 Datsun 710        Datsun 710         22.8
#> 3 Hornet Sportabout Hornet Sportabout  18.7
#> 4 Duster 360        Duster 360         14.3
#> 5 Merc 240D         Merc   240D        24.4
#> 6 Merc 230          Merc   230         22.8

# How can I make the y-axis text color reflect the make of each model?
mtcars %>%
  ggplot(aes(x = mpg, y = model)) +
  geom_point() +
  labs(y = NULL) +
  scale_x_continuous(breaks = seq(0, 35, 5), limits = c(0,35)) +
  theme_bw()

^{Created on 2023-12-06 with reprex v2.0.2}

Here is a similar SO question that uses an example with only two categories (this example has 20; link here)

>Solution :

It’s officially unsupported to have vectorised colors in theme elements, but it does work, so you could do:

mtcars %>%
  ggplot(aes(x = mpg, y = model)) +
  geom_point() +
  labs(y = NULL) +
  scale_x_continuous(breaks = seq(0, 35, 5), limits = c(0,35)) +
  theme_bw() +
  theme(axis.text.y = element_text(
    color = scales::hue_pal()(length(levels(factor(mtcars$make))))[
      as.numeric(factor(mtcars$make))]
      ))

Personally, I would avoid using colors and a legend for axis labels. Why not directly label with facets?

mtcars %>%
  ggplot(aes(x = mpg, y = model)) +
  geom_point() +
  labs(y = NULL) +
  scale_x_continuous(breaks = seq(0, 35, 5), limits = c(0,35)) +
  theme_minimal() +
  theme(panel.spacing.y = unit(0, "mm"),
        strip.text.y.left = element_text(angle = 0),
        strip.background = element_rect(fill = "gray90", color = "white",
                                        linewidth = 1.5),
        strip.placement = "outside") +
  facet_grid(make~., scales = "free_y", space = "free_y", switch = "y")