Variable legend values and colors in GGPLOT

Question

I've been omitting I problem I have regarding the production of maps though different columns. Let's say I have a certain dataset, where every row is a polygon:

> crops <- data.frame(pol_num = c("1", "2", "3","4"),
wheat2020 = c("yes", "yeswater", "yes","no"),
wheat2030 = c("yeswater", "no", "no","no"))

> crops
  pol_num wheat2020 wheat2030
1       1       yes  yeswater
2       2  yeswater        no
3       3       yes        no
4       4        no        no

As you see I have 4 polygons and the suitability of a crop in 2 years. I want to plot these polygons in ggplot with custom colors for each possibility ("yes", "yeswater" and "no").

The only solution I found is to order each column in a certain order:

crops$wheat2020 <- factor(crops$wheat2020, levels = c("yes", "yeswater","no"))

And creat a custom palette with the order of colors I need:

palette <-c("#98E601", "#72DFFE", "#A80100")

And then apply scale_fill_manual. This gets really tricky because each scenario has different possible combinations. In the example you can see how wheat2020 has 3 different values whereas wheat2030 only 2!

The code, simplified, for my plot and a certain scenario is:

ggplot() + 
  geom_sf(data = crops, aes(fill = wheat2020), lwd = 0.001) +
  scale_fill_manual(labels=c("Yes", "Yes water","No"),
                    name="Suitability wheat: 2020",
                    values = palette)

Is there any way to simply indicate which color must correspond each value in the columns, no matter if there is one or two of the possible values missing?

Thanks!

Answer 1

I have created some simple geometry to emulate your data frame better, but essentially all you need is to pass your color values as a named vector:

ggplot() + 
  geom_sf(data = crops, aes(fill = wheat2020), lwd = 0.001) +
  scale_fill_manual(name = "Suitability wheat: 2020",
                    values = c(yes      = "#98E601", 
                               yeswater = "#72DFFE", 
                               no       = "#A80100"))

ggplot() + 
  geom_sf(data = crops, aes(fill = wheat2030), lwd = 0.001) +
  scale_fill_manual(name = "Suitability wheat: 2030",
                    values = c(yes      = "#98E601", 
                               yeswater = "#72DFFE", 
                               no       = "#A80100"))