Parse aesthetic values from the values of the variable

Question

I have data set that has different colors in a variable, my_color.

library(tidyverse)

set.seed(1234)

dat <- tibble(x = 1:5,
              y = 1:5,
              my_color = sample(colors(), 5))

dat
#> # A tibble: 5 × 3
#>       x     y my_color  
#>   <int> <int> <chr>     
#> 1     1     1 grey23    
#> 2     2     2 darksalmon
#> 3     3     3 tan3      
#> 4     4     4 violetred4
#> 5     5     5 lightblue1

I want to plot the points, and each point to be colored according to the color in my_color. I can a color aesthetic using aes(), but the colors are auto-assigned by ggplot2. This is normally a good thing, but in this instance I have specific values I want.

ggplot(dat, aes(x = x, y = y, color = my_color)) + 
  geom_point(size = 3)

My only work around currently is to manually assign each color to itself, e.g.,

ggplot(dat, aes(x = x, y = y, color = my_color)) + 
  geom_point(size = 3) +
  scale_color_manual(values = c("darksalmon" = "darksalmon",
                                "grey23" = "grey23",
                                "lightblue1" = "lightblue1",
                                "tan3" = "tan3",
                                "violetred4" = "violetred4"))

This is not super feasible for a real data set with many more than 5 rows. Is there a way to parse aesthetic values from a given variable directly?

^{Created on 2022-07-26 by the reprex package (v2.0.1)}

Answer 1

One way to solve this is to use fct_inorder from forcats package (it is in tidyverse). This will hinder ggplot to sort alphabetically.

Then we can use scale_colour_manual as usual by assigning the ordered colors:

library(tidyverse)

ggplot(dat, aes(x = x, y = y, color = fct_inorder(my_color))) + 
  geom_point(size = 3) +
  scale_color_manual(values = dat$my_color)