Draw heatmap tiles for all combination of x-y

Question

I asked a question about the heatmap which was solved here: custom colored heatmap of categorical variables. I defined my scale_fill_manual for all combinations as suggested in the accepted answer.

Based on this question, I would like to know how to tell ggplot2 to plot a heatmap with all combination of variables and not just the ones that are available in the dataframe (given that they are already in the scale_fill_manual but are not showing in the final plot).

How can I do this?

The current plotting code:

df <- data.frame(X = LETTERS[1:3], 
                 Likelihood = c("Almost Certain","Likely","Possible"), 
                 Impact = c("Catastrophic", "Major","Moderate"), 
                 stringsAsFactors = FALSE)
df$color <- paste0(df$Likelihood,"-",df$Impact)

ggplot(df, aes(Impact, Likelihood)) + geom_tile(aes(fill = color),colour = "white") + geom_text(aes(label=X)) +
  scale_fill_manual(values = c("Almost Certain-Catastrophic" = "red","Likely-Major" = "yellow","Possible-Moderate" = "blue"))

scale_fill_manual contains all combination of Impact, Likelihood with their respective colors.

so you basically want all remaining tiles, for which there are no values, to be the same color (e.g., grey?) — timfaber, Sep 14 '17 at 13:47
Could you add your current code to the question? One solution might be to expand the dataset to encompass all combinations of `Likelihood` and `Impact` prior to plotting with, e.g., `tidyr::complete`. — aosmith, Sep 14 '17 at 15:02
no not the same color, basically based on the manual fill scale I provide for each combination @timfaber — ifreak, Sep 15 '17 at 06:38
I don't see a fill color for every possible combination of `Likelihood` and `Impact`. What should the other fill colors be? — aosmith, Sep 15 '17 at 13:37
@aosmith yes, its because the list is too long to put here. But there is a color for each combination — ifreak, Sep 15 '17 at 14:52
In that case, my idea above should work. You can expand your dataset so all combinations of the two variables are present prior to making the `fill` variable and plotting. That task of expanding the dataset is fairly straightforward with`tidyr::complete`. — aosmith, Sep 15 '17 at 15:17

score 0 · Answer 1 · answered Sep 15 '17 at 19:30

Similar to @aosmith I tried expand.grid to get a finite set of combinations but tidyr::complete() works pretty nice as well. Add the colors and letters and fill using a set color range.

df <- data.frame(Likelihood = c("Almost Certain","Likely","Possible"), 
                 Impact = c("Catastrophic", "Major","Moderate"), 
                 stringsAsFactors = FALSE)
df2 <- df %>% tidyr::complete(Likelihood,Impact) # alt expand.grid(df)

df2$X <- LETTERS[1:9] # Add letters here
df2$color <- paste0(df2$Likelihood,"-",df2$Impact) # Add colors


ggplot(df2, aes(Impact, Likelihood)) + geom_tile(aes(fill = color),colour = "white") + geom_text(aes(label=X)) +
  scale_fill_manual(values = RColorBrewer::brewer.pal(9,"Pastel1"))

Draw heatmap tiles for all combination of x-y

1 Answers1