CODEWITHSUNDEEP
rggplot2themes
Paula
Paula

Reputation: 703

Creating ggplot theme: custom colors for variable levels

I'm trying to create a new theme in ggplot2 for plotting political parties for my country, I want to set the color for some and other like random_party_1 y random_party_2 set a random color.

data <- data.frame(party = c('FA','PN', 'random_party1','random_party2', 'FA','PN', 'random_party1','random_party2'), 
                   year = c(2010, 2010, 2010, 2010, 2008, 2008, 2008, 2008), 
                   value = c(50, 20, 30, 10, 50, 50, NA, NA), 
                   color = c('#013197','#B0C2D3','assign_random_color1','assign_random_color2','#013197','#B0C2D3','assign_random_color','assign_random_color'))

Then I want to use this as a fill or color argument but I don't know how to implement it without declaring it manually like because it's going to depend on the data of each year:

cols <- c("FA" = "#013197", "PN" = "#B0C2D3", "random_party1" = "assign_random_color1", "random_party2" = "assign_random_color2")
p + scale_colour_manual(values = cols)

Upvotes: 0

Views: 229

Answers (1)

Allan Cameron
Allan Cameron

Reputation: 174586

I think the best way to achieve this is via a custom function:

colorize <- function(vec)
{
  levs <- if(is.factor(vec)) levels(vec) else levels(factor(vec))
  predefined <- c("FA", "PN")
  pal <- c('#013197','#B0C2D3')
  pal <- pal[match(levs, predefined)]
  blanks <- which(is.na(pal))
  pal[blanks] <- sample(colours(100), length(blanks))
  pal
}

So suppose we set up your plot like this:

p <- ggplot(data, aes(factor(year), value, fill = party)) + geom_col(position = "dodge")

Then we can run our code once to get fixed colors for the predefined parties and random colors for the unassigned parties:

p + scale_fill_manual(values =  colorize(data$party))

enter image description here

And we can see only the random parties' colors change if we run exactly the same code again:

p + scale_fill_manual(values =  colorize(data$party))

enter image description here

Created on 2020-06-11 by the reprex package (v0.3.0)

Upvotes: 1

Related Questions