Is it possible to make it show within-group densities instead of counts?
library(ggplot2);data(diamonds)
ggplot(diamonds, aes(carat, depth)) +
stat_bin2d(bins=40)+ facet_wrap(~color)
This would make it easier to compare patterns among the groups, as some groups may naturally be more occurring.
The question is slightly similar to: How to scale (normalise) values of ggplot2 stat_bin2d within each column (by X axis) which is also lacking an answer.
ggplot(diamonds, aes(carat, depth)) +
stat_bin2d(bins=40, aes(fill = ..density..))+ facet_wrap(~color)
Or would you be happy with a kernel density estimate?
ggplot(diamonds, aes(carat, depth)) +
stat_density2d(aes(fill = ..density..), geom = "tile", contour = FALSE, n = 25) +
facet_wrap(~color) +
scale_fill_gradient(low = "light blue", high = "dark red")
Or with default grid:
ggplot(diamonds, aes(carat, depth)) +
stat_density2d(aes(fill = ..density..), geom = "tile", contour = FALSE) +
facet_wrap(~color) +
scale_fill_gradient(low = "light blue", high = "dark red")