I have been a big fan of Hadley Wickham’s ggplot package for a while. This last week I was building a shiny app and wanted to find out exactly which colours were being used in the plot. After a bit of googling I ended up at this stackoverflow post, which led me to the ggplot_build function which I found seriously useful.

A little demo:

library(ggplot2)

# create a sample plot
p = ggplot(mtcars, aes(x = hp, y = mpg, colour = factor(cyl), shape = factor(am))) +
  geom_point(size = 3) +
  labs(colour = "No. Cylinders", shape = "Auto/Manual",
       x = "Horsepower", y = "Miles per Gallon",
       title = "Scatterplot of Horsepower vs Miles per Gallon")
# and display it
p

One way we can look at what info ggplot is using to make the plot is with the ggplot_build() function. This outputs a list of, amongst other things, a data frame of all the aesthetics.

# we can use the ggplot_build function to have a look at exactly what information ggplot is using
# to construct the plot
ggplot_build(p)$data

## [[1]]
##     colour shape   x    y PANEL group size fill alpha stroke
## 1  #00BA38    17 110 21.0     1     4    3   NA    NA    0.5
## 2  #00BA38    17 110 21.0     1     4    3   NA    NA    0.5
## 3  #F8766D    17  93 22.8     1     2    3   NA    NA    0.5
## 4  #00BA38    16 110 21.4     1     3    3   NA    NA    0.5
## 5  #619CFF    16 175 18.7     1     5    3   NA    NA    0.5
## 6  #00BA38    16 105 18.1     1     3    3   NA    NA    0.5
## 7  #619CFF    16 245 14.3     1     5    3   NA    NA    0.5
## 8  #F8766D    16  62 24.4     1     1    3   NA    NA    0.5
## 9  #F8766D    16  95 22.8     1     1    3   NA    NA    0.5
## 10 #00BA38    16 123 19.2     1     3    3   NA    NA    0.5
## 11 #00BA38    16 123 17.8     1     3    3   NA    NA    0.5
## 12 #619CFF    16 180 16.4     1     5    3   NA    NA    0.5
## 13 #619CFF    16 180 17.3     1     5    3   NA    NA    0.5
## 14 #619CFF    16 180 15.2     1     5    3   NA    NA    0.5
## 15 #619CFF    16 205 10.4     1     5    3   NA    NA    0.5
## 16 #619CFF    16 215 10.4     1     5    3   NA    NA    0.5
## 17 #619CFF    16 230 14.7     1     5    3   NA    NA    0.5
## 18 #F8766D    17  66 32.4     1     2    3   NA    NA    0.5
## 19 #F8766D    17  52 30.4     1     2    3   NA    NA    0.5
## 20 #F8766D    17  65 33.9     1     2    3   NA    NA    0.5
## 21 #F8766D    16  97 21.5     1     1    3   NA    NA    0.5
## 22 #619CFF    16 150 15.5     1     5    3   NA    NA    0.5
## 23 #619CFF    16 150 15.2     1     5    3   NA    NA    0.5
## 24 #619CFF    16 245 13.3     1     5    3   NA    NA    0.5
## 25 #619CFF    16 175 19.2     1     5    3   NA    NA    0.5
## 26 #F8766D    17  66 27.3     1     2    3   NA    NA    0.5
## 27 #F8766D    17  91 26.0     1     2    3   NA    NA    0.5
## 28 #F8766D    17 113 30.4     1     2    3   NA    NA    0.5
## 29 #619CFF    17 264 15.8     1     6    3   NA    NA    0.5
## 30 #00BA38    17 175 19.7     1     4    3   NA    NA    0.5
## 31 #619CFF    17 335 15.0     1     6    3   NA    NA    0.5
## 32 #F8766D    17 109 21.4     1     2    3   NA    NA    0.5

I found this very useful to double check that all the things I was expecting to happen, were actually happening and in to solve my particular problem of finding out exactly which colours I was using. For more detail on all the information ggplot_build can provide have a look at the docs.