R 中具有分类变量的热图

Question

I am trying to create a heatmap with long form data to assess the correlation between membership of individual groups on specific outcome measure scores. I have converted the data to long format as shown. Some individuals may be members of more than 1 group.

df1 <- data.frame(ID = c(1, 2, 3, 4, 5),
               Group1= c(1, 1, 0, 1, 0),
               Group2 = c(1, 0, 1, 0, 0),
               Measurement = c("Pain", "Pain", "Fatigue", "Stiffness", "Fatigue"),
               Score = c(2, 3, 3, 2, 3))

df1
ID Group1 Group2 Measurement Score
1  1      1      1        Pain     2
2  2      1      0        Pain     3
3  3      0      1     Fatigue     3
4  4      1      0   Stiffness     2
5  5      0      0     Fatigue     3

I am sure this is a very basic question so my apologies. I would like to create a heatmap-like plot in ggplot with a colour scale representing the scores and x axis= Group memberships, y-axis = Measurements

Running the code below yields the following:

fig1 <- ggplot(df1,  aes(x =  c("Group1", "Group2"),
                     y = "Measurement", 
                     fill = "Score")) + geom_tile()

Error: Aesthetics must be either length 1 or the same as the data (5): x

Is there a way of having a list of categorical variables on the y and x axes as described?

Thanks

Answer 1

There are two issues with your code:

1. You use quotes " around column names which is the reason why you get the error message.
2. You can't map two columns on one aesthetic, ie even when removing the quotes c(Group1, Group2) will not work. Instead combine both variables into one for which I use dplyr::case_when .

library(ggplot2)
library(dplyr)

df1 <- df1 %>%
  mutate(Group = dplyr::case_when(
    Group1 == 1 & Group2 == 1 ~ "Group 1 & 2",
    Group1 == 1 ~ "Group 1",
    Group2 == 1 ~ "Group 2",
    TRUE ~ "None"
  ))

ggplot(df1, aes(
  x = Group,
  y = Measurement,
  fill = Score
)) +
  geom_tile()