Kernel Density (2D)

I use so many KDE’s that you might be tempted to think I am employed as a PR rep by the Gauss estate. Here are some examples of 2 dimensional non-spatial KDE visualizations.

An image from the third chapter of my dissertation looking at transit travel time estimates. This plot summarizes the relative difference (new method) in estimated travel times to the baseline (old method). The vertical scale is logged to make visual comparison possible, and curved lines represent differences in absolute values. The KDE summarizes more than 100 million point estimates. The visualization was done in base R.
This is actually a 1 dimensional radial KDE, showing geographic direction relative to a central place.