I am currently working on a research project at UCSC and have built a tool similar to this:
The application I am working on is different, though, in that it allows the user to specify an address and a radius.
During testing, I have noticed that when a user specifies a larger radius, results become more sparse.
For example, a search for Santa Cruz (which resolves to 36.9741171, -122.0307963) with a 1km radius yields 96 results
(count == 100) while the same query with a 10km radius only yields 15 results.
The same query with a 5km radius yields 14 results.
Among the 3 cases, a few results overlap but not as many as you would expect.
Can anyone explain this phenomenon?