0

Is there a data entity that would correspond to population clusters within a zip code tabulation area? Something that would correspond, even if only roughly, with a neighborhood or housing tract or section of a town?

I understand that the zipcode defines a postal route and that the zipcode tabulation area was a concession to the lay people who rely upon zipcodes and that with my question it's like I'm asking for a fish sandwich at a pizza shop. But we're constantly being sent company employee databases with home zipcodes only (not full home addresses) and we have to do an analysis to see how well their employee population is served by our provider network. How far do they have to travel to see a particular kind of provider? So we need some way (in the absence of street addresses) to identify where in a given zipcode people are likely to live, so we can compute distances to provider office locations from non-random starting points other than the zipcode centroid.

5
  • The US Census Bureau uses census blocks as the smallest geographical unit in their surveys. Depending on population density, you may find that blocks are smaller than zip code zones. You could intersect zip codes with census blocks, but I doubt they would match well, since zip code zones are not based on blocks. You also have the issue that you are receiving data for the potentially larger unit (zip code zone), and are trying to delineate it smaller without the data to break it up. You might try using a heat map of population density of census blocks to see hot spots in a zip code zone.
    – Baltok
    Dec 7, 2022 at 23:11
  • Thanks for the suggestions. We could use the state and county attributes to find all of the blocks in the same state/county as the zipcode and then identify any of them that are fully contained within the zipcode. That might work much of the time. We could rank the blocks by their population and then allocate employees to the blocks accordingly. There's a significant margin of error but it would be more defensible than putting them all at the centroid.
    – Tim
    Dec 8, 2022 at 11:04
  • What would be ideal is for these companies to pre-process their home-addresses database and send us the geocoordinates of the street intersection nearest to the employee address. That approach would obscure the employee's precise location and give us real data to work with.
    – Tim
    Dec 8, 2022 at 11:04
  • 1
    Are the companies willing to do that, though? That sounds like a bit of work on their end. If they aren't willing, then reducing the granularity without more/better data will be tough, and of questionable accuracy. There's a reason the gis community jokes about having BAD (Best Available Data) data.
    – Baltok
    Dec 9, 2022 at 14:36
  • It's baffling to me how a product using such a specious algorithm has managed to become the de facto standard.
    – Tim
    Dec 9, 2022 at 15:50

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Browse other questions tagged or ask your own question.