Be wary that correlation does not imply causation... but with that being said, love seeing your continued analysis on this and look forward to seeing how it pans out this time!

Do other exchanges show users online? Be very cool if you could use that to get a larger sample of data, although I think Polo is the leader so probably wouldn't be too much deviation.

Also be interesting to know geographically where the users are. Obviously not something that can be easily captured, but I've always been surprised that there seem to be so many more people on early morning as opposed to evening time (I'm in US Pacific time)