It's a slow process (as I've been respectful of the rate-limit on the reverse geocoding API I'm using), but I finished mapping four states and hope to complete the other 46 by the end of this week.
I went and looked, before reading your second paragraph, and the clusters immediately jumped out, and it either seemed suspect or wrong. Then I had to come back here to see why it was wrong.
Examples:
All the ones in Ontario, Canada, e.g. Fullmer Construction, are really in Ontario, California.
The Christian Valley School on Artesia boulevard in Alberta was picked up instead of the one in Artesia, CA.
Simpson Labs is geocoded to Valencia, Spain instead of Valencia, CA.
GW Lawrence on Wilson street in Cordoba, Argentina instead of Cordoba, AR.
Is it possible to add a ggogle maps business name listing based on address next to the reported company name?
Cool project though.
You're right that it would be interesting to do analysis on population-adjusted loan amounts for different regions... maybe something like a loan $ per capita map?
What I think would be much more interesting is to see what businesses have physical B&M locations. In other words, cross reference Yelp/Google Places with the legal entity's address. Or maybe cross referenced against unemployment numbers by state/city (if that data exists).
Nit: what you're doing is geocoding, not reverse geocoding. Geocoding is the process of turning strings into lats/longs. Reverse is the reverse of that, namely lats/longs into strings.