In a paper published in PLOS One, Garrett Dash Nelson and Alasdair Rae explore whether megaregions—i.e., a region centred on a major metropolitan area—can be determined algorithmically, using commuter flow data. In the end they conclude that “any division of space into unit areas will have to take into account a ‘common sense’ interpretation of the validity and cohesion of the regions resulting from an algorithmic approach. For this reason, the visual heuristic method coupled with the algorithmic method offers a good combination of human interpretation and statistical precision.” In the process, they’ve generated a series of maps that are fascinating on several levels, including a final map of megaregions that combines algorithmic results with visual heuristics (i.e., human judgment). [Atlas Obscura]