Actual result of ..ndensity.. does not match the documentation

[timgoodman]

In help(stat_bin), it defines ..ncount.. as "count, scaled to maximum of 1" and ..ndensity.. as "density, scaled to maximum of 1". From this I would expect that:

..ncount.. is ..count.. / max(..count..)

and that

..ndensity.. is ..density.. / max(..density..)

However, the latter is not the case. Instead, ..ndensity.. appears to be ..count.. / max(..density..)

If this is a bug, it looks like it could be fixed with a one line change to R/bin.R (which I'd be happy to submit as a pull request if you want). But if this is the intended behavior, perhaps the documentation could be made more clear.

library(ggplot2)

standard_normal <- data.frame(X = rnorm(1E5, 0, 1))

# ..ndensity.. is ..count.. / max(..density..)
# which is not "density, scaled to a maximum of 1"
ggplot(standard_normal, aes(x = X)) +
  geom_freqpoly(aes(y = ..ndensity..), binwidth = 0.2, center = 0, size = 1.5, color = 'black') +
  geom_freqpoly(aes(y = ..count../max(..density..)), binwidth = 0.2, center = 0, size = 0, color = 'cyan')

[timgoodman]

In particular, the changes to R/stat-bin.r in this commit included removing the line:

results$ndensity <- results$density / max(abs(results$density), na.rm = TRUE) 

while adding the line:

ndensity = count / max(abs(density)),

I assume the change from density to count in the numerator was unintentional. The new code was later moved from R/stat-bin.r to R/bin.r in this commit.