Michiel de Hoon wrote:
Thanks!
I uploaded a demo on sourceforge:http://sourceforge.net/tracker/index.php?func=detail&aid=1952339&group_id=80706&atid=560722
Michiel,
I have committed a modification of the demo, and also made some tweaks in hexbin itself. Manuel Metz also made some changes and additions, and more may be in order as the function is tested and used by more people. The main change I made is to use log10 for the log scaling. This the usual expectation, I believe.
For interpreting the actual value of what is plotted--the counts per cell--it might be good to output the area of the cell, or have an option to plot something like counts per unit area divided by total counts, which would be a probability density.
Thanks again for contributing hexbin.
Eric