Changing colour of existing geom, with no grouping

user_10782538 0
Viewed 81

This is likely just a blunder from my side, not remembering how to do this. Non-the-less. Imagine a scenario where you have a function generating a ggplot. You give no grouping, and the points are all coloured black. For simplicity, imagine the plot below is the one returned.

library(ggplot2)
data(mtcars)
p <- ggplot(mtcars, aes(y = mpg, x = hp)) + 
      geom_point()

How would I now change the colour of the points in p, without modifying the function that generated the plot itself. Eg. changing the code to geom_point(col = "green") would not be the an option.

ggplot2 r
2 Answers

An easy way to do it is to add a dummy aesthetic mapping after the fact:

 p + aes(color = "forestgreen") + scale_color_identity(guide = guide_none())

enter image description here

A more permanent alternative, which is the equivalent of going back and writing the color = "forestgreen" parameter back into the geom_point call is:

p$layers[[1]]$aes_params <- list(colour = "forestgreen")
user_125003150

One approach is to create a ggplot_build object, modify it directly, and rebuild the ggplot object:

p <- ggplot(mtcars, aes(y = mpg, x = hp)) +
       geom_point()
q <- ggplot_build(p)
q$data[[1]]$colour <- "cornflowerblue"
p <- ggplot_gtable(q)
plot(p)

enter image description here

user_130953260
Related