Plot a heat map with a gradient for the tiles based on the proximity of a specific area of the plot

Question

My goal is to use a heat map to plot the data below. I have obtained the layout of the heat map I wanted, but I am struggling to find a way to color the tiles as intended. My guess is that the code below colors the tiles proportional to the number inside the tile itself. However, this is not what I need. I would like to:

1. have the same color within each of the rectangles (yellow for the rectangle on the left and blue for the rectangle on the right);
2. a gradient from yellow to blue for the other tiles proportional to how far away the tiles are from the blue area.

Thanks to anyone who will help!

library(tidyverse)

# First, I create the simulated dataset with 200 individuals
set.seed(1243)  # set seed for reproducibility

#I simulate test 1 scores
test1_score <- sample(c(3.5, 3.8, 4), 200, replace = TRUE)

#I simulate test 2 scores
test2_score <- round(runif(200, 0, 38))

#I create diagnostic classes based on test 2 scores
test2_class <- ifelse(test2_score < 20, "non meet",
                      ifelse(test2_score < 30, "below 40th",
                             ifelse(test2_score < 35, "above 40th",
                                    ifelse(test2_score < 37, "meet", "master"))))

#I create exit_program variable based on test 1 scorea and test 2 classes
exit_program <- ifelse(test1_score == 4 & test2_class %in% c("above 40th", "meet", "master"), 1, 0)

#I create data frame with simulated data
df <- data.frame(ID = 1:200, test1_score, test2_class, exit_program)

#I calculate the group size for each combination of test 1 scores and test 2 classes
df2 <- df |>
  group_by(test1_score, test2_class, exit_program) |> 
  summarize(n = n()) |> 
  mutate(test2_class = factor(test2_class, levels = c("non meet", "below 40th", "above 40th", "meet", "master"))) |> arrange(test2_class)

#plot data
df2 |> 
  ggplot(aes(x = test2_class, y = as.factor(test1_score))) +
  theme(legend.position = "none", panel.background = element_rect(fill = "white"),
        axis.text = element_text(size = 12), axis.title = element_text(size = 14)) +
  geom_tile(aes(fill = n)) + 
  geom_text(aes(label = n), size = 7, family = "sans") +
  labs(x = "classification", y = "scores", size = 10) +
  scale_fill_gradient(low = "#56B4E9", high = "#F0E442", guide = "none") +
  geom_rect(aes(xmin = 2.5, xmax = 5.5, ymin = 2.5, ymax = 3.5), linewidth = 2, color = "#0072B2", fill = NA) +
  geom_rect(aes(xmin = 0.5, xmax = 1.5, ymin = 0.5, ymax = 3.5), linewidth = 2, color = "#E69F00", fill = NA)

Answer 1

Maybe this is what you are looking for. You are right. Mapping n on fill means to color the tiles by the value of n.

As I understand the question you want each "column" to have the same color and colored with a gradient from yellow on the left to blue on the right. To this end you could add a column with a measure of the distance from the left or the right end to your data which could then be mapped on the fill aes. One option would be to convert your test2_class to a numeric then compute the absolute distance from the maximum value (which corresponds to the "master" level).

library(tidyverse)

df2 <- df2 |>
  ungroup() |>
  mutate(
    fill = as.numeric(test2_class),
    fill = abs(fill - max(fill))
  )

#plot data
df2 |> 
  ggplot(aes(x = test2_class, y = as.factor(test1_score))) +
  theme(legend.position = "none", panel.background = element_rect(fill = "white"),
        axis.text = element_text(size = 12), axis.title = element_text(size = 14)) +
  geom_tile(aes(fill = fill)) + 
  geom_text(aes(label = n), size = 7, family = "sans") +
  labs(x = "classification", y = "scores", size = 10) +
  scale_fill_gradient(low = "#56B4E9", high = "#F0E442", guide = "none") +
  geom_rect(aes(xmin = 2.5, xmax = 5.5, ymin = 2.5, ymax = 3.5), linewidth = 2, color = "#0072B2", fill = NA) +
  geom_rect(aes(xmin = 0.5, xmax = 1.5, ymin = 0.5, ymax = 3.5), linewidth = 2, color = "#E69F00", fill = NA)