No way to specify number of bins in stat_density_2d

[ecoRoland]

stat_density2d used to pass the ellipses to stat_contour. That way one could specify the number of bins.

Pre-v2.0.0:

m <- ggplot(faithful, aes(x = eruptions, y = waiting)) +
  geom_point() +
  xlim(0.5, 6) +
  ylim(40, 110)
m + geom_density2d(bins = 3)

Since stat_density_2d doesn't do that anymore, I don't see a way to specify the number of bins. If I'm not missing something that seems like a regression.

[has2k1]

That is maybe an inconsistency (given that bins is used for partitioning the grid in other cases) in the API. Use n = 3.

[ecoRoland]

@has2k1 I don't follow. n = 3 is passed to kde2d and does not specify the number of bins.

[has2k1]

@ecoRoland, maybe I do not understand what you mean by bins, but n does partition the grid and that is equivalent to bins. In 2nd example from the bottom at http://docs.ggplot2.org/dev/geom_density_2d.html, that is n=100 which is 100 bins.

[MikeWise2718]

Can you replicate this http://stackoverflow.com/questions/19329318/how-to-correctly-interpret-ggplots-stat-density2d effect with the n=4 parameter? I couldn't...

[has2k1]

Oh. Okay, you cannot do that anymore. The 2.0 clean up means parameters do not fall through any more. I'll submit a PR.