2D Kernel density of class ratio?

I have a dataset with lat/long coordinates and a binary class of interest.

# Simulating the dataset
df <- tibble(x = rnorm(100),
             y = rnorm(100),
             class = rbernoulli(100))

Now, my objective is to visualize how the rate of class == 1 varies in the 2D space defined by x and y. I'm aware of:

# Density of points
ggplot(df) + 
  geom_density_2d(aes(x = x, y = y))

But I would need to do the same plot with the percentage of class == 1. instead of the density of points itself. It would be basically the ratio of density corresponding to class == 1 by the density corresponding to the whole dataset.

Is there a way to maybe tweak stat_density_2d() in order to achieve that goal? Or alternatively, is there any method to obtain this plot?

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.