pygtc
pygtc copied to clipboard
order of operations in __plot2D
I see in __plot2D
that the Gaussian smoothing is done after normalizing and finding the contour levels to show. Seems like it would be more consistent to use the smoothed density estimate throughout. In particular, if the smoothing is not conservative, the contour levels may be thrown off noticeably at ~2 sigma and higher.
Thanks Adam, that is a very sensible suggestion. In hindsight I guess there's no good reason to have done the inverse. Let me look into this!