Hate to say it, but I'm not sure this type of visualization is ideal for this data. Given that the actual clustering is meaningless, I would think a simple histogram would expose the relationship between the different cities better.<p>My mind could be changed if this exposed a hierarchy, for example if the US was a single bubble which could be clicked-into. Or if there was a meaning behind the clustering, like if geographic proximity determined how close the bubbles were.
I made a little map visualization using the great Datawrapper tool on the front page:<p><a href="https://datawrapper.dwcdn.net/TSri6/3/" rel="nofollow">https://datawrapper.dwcdn.net/TSri6/3/</a><p>note: there are a few inaccuracies. 7 people from "Vatican City" weren't counted because it wasn't recognized as a country. 3-4 people listing "Remote" or "Nomad" each were left of the map. I summed the counts from US states into one number using my poor JS skills, hopefully i got that right. Of course there may be other mistakes I don't know about :)