If we know the CBGs that make up Rocky National Park (for example), could we use Neighborhood Patterns to estimate traffic?

Questions & Help
1

Starting a new thread for an idea related to this thread.

I have an idea for estimating large park traffic that I want to get some feedback on. If we know the CBGs that make up Rocky National Park (for example), could we use Neighborhood Patterns to estimate traffic? Neighborhood Patterns has visits/stops data aggregated at the CBG level, instead of POI-level. One challenge would be to account for devices that are pinging in multiple CBGs in the park, because they would show up multiple times in the Rocky National Part Neighborhood patterns. I think we could mitigate some of the duplicated counts with the device_home_areas and device_daytime_areas
columns. Any thoughts?

cc: @Nick_H-K_Seattle_University @Andrew_Bailey_UT_Chattanooga @Jude_Bayham_Colorado_State_U

2

Interesting approach. I think that has some potential for large parks like RMNP. Yes, the double counting is a caveat, but it is for my approach using pois within the park as well. Spot checking RMNP certain cbgs cross the park boundaries, so you may introduce some commission error around the border with the nearby town. However, I think that error is smaller than the omission error of just using pois within the park.

3

That seems reasonable to me. How many parks cross multiple CBGs? CBGs are pretty big in rural areas no?

4

I would be interested to see how well this works. I won’t have time to test it out myself for a while. If anyone else tests it out, I’d be really interested to see the methods and results! Please keep me updated