rggplot2geom

Dynamic Scaling of geom_tile in ggplot2

I have a data frame with the following structure:

  x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 x11 x12 proportion index cumulative_proportion
1  0 -1  0  0  0 -1  1  1  0   1   0   0    0.28423     1               0.28423
2  0 -1  0  0  0 -1  1  1  1   1   0   0    0.20511     2               0.48934
3  1 -1  0  0  0 -1  1  1  0   1   0   0    0.05751     3               0.54685
4  0 -1  0  0  1 -1  1  1  0   1   0   0    0.02388     4               0.57073
5  0 -1  0  0  0 -1  1  1  0   1   1   0    0.02217     5               0.59290
6  0 -1  0  0  0 -1  1  1  0   1   0   1    0.02098     6               0.61388

I'd like to do a visualization using geom_tile(), where the var variables are on the y-axis, and the x-axis is based on the categorical index, with the filling determined by whether the variables are -1, 0, or 1. The issue is that at the same time, the x-axis should show the cumulative_proportion variable and the tiles should be scaled based on proportion. Here's a non-ggplot example, some code that constructs a non-scaled version and the example data frame:

example

plot.df <- reshape2::melt(plot.df, id.vars=c("index", "proportion", "cumulative_proportion"))
ggplot(plot.df, aes(x=index, y=variable, fill=as.factor(value))) +
  geom_tile()

example2

structure(list(x1 = c(0, 0, 1, 0, 0, 0, 1, 0, 0, 1), x2 = c(-1, 
-1, -1, -1, -1, -1, -1, -1, -1, -1), x3 = c(0, 0, 0, 0, 0, 0, 
0, 0, 0, 1), x4 = c(0, 0, 0, 0, 0, 0, 0, 1, 1, 0), x5 = c(0, 
0, 0, 1, 0, 0, 0, 0, 0, 0), x6 = c(-1, -1, -1, -1, -1, -1, -1, 
-1, -1, -1), x7 = c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1), x8 = c(1, 
1, 1, 1, 1, 1, 1, 1, 1, 1), x9 = c(0, 1, 0, 0, 0, 0, 0, 0, 0, 
0), x10 = c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1), x11 = c(0, 0, 0, 0, 
1, 0, 0, 0, 0, 0), x12 = c(0, 0, 0, 0, 0, 1, 1, 0, 1, 0), proportion = c(0.28423, 
0.20511, 0.05751, 0.02388, 0.02217, 0.02098, 0.0165, 0.01383, 
0.01185, 0.01013), index = 1:10, cumulative_proportion = c(0.28423, 
0.48934, 0.54685, 0.57073, 0.5929, 0.61388, 0.63038, 0.64421, 
0.65606, 0.66619)), row.names = c(NA, -10L), class = "data.frame")
Solution

  • One option would be to switch to geom_rect which requires some additional data wrangling to compute or set the positions of the rectangles:

    library(ggplot2)
library(dplyr, warn = FALSE)

plot.df <- plot.df |>
  mutate(
    ymin = as.numeric(factor(variable)) - .5,
    ymax = as.numeric(factor(variable)) + .5
  ) |>
  mutate(
    xmin = lag(cumulative_proportion, default = 0),
    xmax = cumulative_proportion,
    .by = variable
  )

ggplot(plot.df, aes(y = variable, fill = as.factor(value))) +
  scale_y_discrete() +
  geom_rect(aes(xmin = xmin, xmax = xmax, ymin = ymin, ymax = ymax), color = "white")