LKremer/ggpointdensity

Would be nice to color points using 2D kernel density

DarwinAwardWinner opened this issue · 2 comments

Rather than counting nearest neighbors, it might be nice to be able to color each point based on the interpolated value of the 2D kernel density estimate at that point's location. You can do this using fields::interp.surface.

This also potentially helps with #2, since I'm pretty sure 2D KDE's runtime isn't quadratic in the number of input points.

I added method="kde2d"to the geom and stat. It is indeed faster for many points (~ >20k) but slower for few points.