Yeah, the data geek in me instantly realized we had to add a time axis, and then a lag, and then variables to incorporate the actual diversities within, and variance in the time lag as well as controls for the aging and . . . many many variables. I’d need more data than anybody actually has, scores of experts optimizing it, and some pretty impressive processing power to do a first run, and then we’d have to another reality check. . . a few dozen cycles of that, and all with the likelihood that the ‘America’ vs ‘Canada’ vs ‘Korea’ (all Korea? Are Minnesota and Texas different?) are dropped because they carry no real weight in the analysis.
Too much work with no real expectation of success, let’s just focus on individual people, right? 
