Hi Sander,
thank you. Yes, I am ware of GeoHistogram, and I have used it before, e.g. here. I often don't get it to look exactly as I want it, particularly if the density of points is low.
For distributions I usually use, well a distribution rather than a histogram, like here. I have used that for the surface of a sphere, too. There are some nice figures for the representation in random matrix theory with that sort of distribution. I look at certain transition/probability matrices, i.e. with largest eigenvalue 1, so that they are on the surface of a sphere. Things like that have come up in models I developed in population matrix theory.
I do, however, agree that GeoHistogram is a very useful function.
I also think that I haven't even begun to properly analyse the GTD, I wanted to try some machine learning on it, now that GPUs are supported again on OSX. It would also be cool to combine the data here with what can be found in the GDELT database from my other post. GDELT also contains data on acts of terrorism, but complementary information. I'd love to see what can be learnt from that.
Thank you for your comment!
Cheers,
Marco