I hope that makes sense-- it's a bit difficult to word. We all know that more stuff happens in the places with more people, and less stuff happens in places with less people. What are some strategies for factoring out population and population density to check if higher or lower rates of a data point are actually significant?
For example, my supervisor gave me a set of geofences around various hospitals on the east coast. They have a field for how many unique cargo truck stops (derived from GPS pings) happened at each hospital for the amount of time that was measured. I'm supposed to analyze the data for patterns in which hospital attributes led to higher numbers of stops. Obviously, the most visited hospitals were in population centers, and the least visited hospitals were rural. This isn't really an interesting or meaningful finding, but it's the only pattern there was.
This leaves me confused as to how I should tell which hospitals actually received significant amounts of traffic once population is accounted for and controlled out.
Is there a standardized procedure for doing this? I feel like there would be, and I just haven't learned about it yet. The only thing I can think of would be to make a field of "visits per 1,000 in population" or something like that, but I don't know if that's an appropriate way to handle the data.
