Neat bit of data!
I'd probably approach this differently, myself. You're looking for location data matching certain locations which is fine, but I think a more comprehensive way to capture this would be to scrape the location data out and then using a script, match those users to countries and throw away any non-location data that is present.
Leveraging the google maps API you could pretty easily get country from pretty much any scrap of location data, or you could manually match through a list (or separate db table) of country/city locations.
Great work though regardless, super interesting to see some of these numbers! Fun little project for sure!
This was a pretty simple script to test the idea. The metadata on Hive is really messy, so it needs some manual processing. Given how few active users we have it could be done by a few people taking a region each. I manage the UK list.