Seems like a better approach would be to identify cities with the most startup "clusters"? Identify a cluster, then look at the population(s) of that census tr

Seems like a better approach would be to identify cities with the most startup "clusters"? Identify a cluster, then look at the population(s) of that census tract or zip or whatever. Using city-level population just makes small cities look great (Boulder) and big cities look terrible (NYC), when I'm pretty sure there are neighborhoods in NYC that blow away Boulder.