3

I'm not a GIS professional, it's just that my current project is heavily reliant on geospatial data.

I'm currently working on a project that has to do with transportation behavior. My hypothesis is that the specific behavior I'm interested in has a regional component to it. For example, agents moving through New York might display this behavior more heavily than agents moving through Texas.

What I'd like to do is have the United States partitioned into uniform clusters (imagine if the states were all the same size) such that for any given transportation route that an agent takes, I can track the clusters that it travels through. This linepath can then be used as input into my main model of interest.

I had considered Zip Codes, as they appear to be roughly uniform, but I have a fear that they are too small/too granular for my purposes. Zip Code Regions have the opposite problem: they are too big.

One other constraint I have is that I work with latitudes and longitudes exclusively. So if a system exists that can take a (lat,long) pair and map that to a cluster, that would be the most ideal solution.

2 Answers 2

1

Without knowing much about what you are trying to analyze, it's difficult to say what an appropriate grouping would be. You might want to simply define something for yourself, e.g. 1-degree (or some even division) lat/long blocks if you want to stick with those coordinates and easily classify them. Whatever you use, I'm sure you're aware that can have an impact on the results of any analyses.

A couple of suggestions that you could look into:

  • Particularly if your interest is at all population-related, you could look at the US Census Bureau divisions, although they will tend toward administrative boundaries at larger areas. Or perhaps their Zip Code Tabulation Areas, which are related to but not the same as USPS Zip Code areas? (And at one time there were three-digit areas, which might be a good fit in-between what you've already looked at.) Here's a place to start: https://www.census.gov/programs-surveys/geography/guidance.html
  • Another possibility, if you're committed to lat/long, is the USGS quadrangle boundaries, which are defined based on lines of latitude and longitude. You will most commonly encounter the 7.5-minute series (showing an area of 7.5 minutes of latitude by 7.5 minutes of longitude), but there is also a 15-minute series, 30-minute by 60-minute, and more if you want larger areas. The USGS produced topographic maps of these areas, which I think are unimportant to you, but you should be able to find shapefiles of the boundaries or indexes that relate the latitude/longitude grid to quad names.

I hope that at least gives you some useful areas of exploration. Good luck!

1

Since you are looking at transportation behavior specifically, you could consider CBSAs or the commuting zones (CZs) popularized by Raj Chetty and coauthors in (https://scholar.harvard.edu/files/hendren/files/mobility_geo.pdf). The idea is to identify geographic areas where people tend to circulate together.

They offer this description / procedure:

Commuting zones are geographical aggregations of counties that are similar to metro areas but cover the entire U.S., including rural areas (Tolbert and Sizer 1996)

We map these ZIP codes to counties based on the 1999 Census crosswalk between ZIP codes and counties. We then aggregate counties into Commuting Zones using David Dorn’s county-to-CZ crosswalk (download file E6). The counties in the U.S. Census Bureau crosswalk and in Dorn’s crosswalk are not identical because they correspond to county definitions at different points in time; in particular the U.S. Census Bureau crosswalk includes changes between 1990 and 1999. We make manual adjustments for changes that affected 200 or more people. Using this procedure, we identify the CZ of 38,839 ZIP codes.

Don't write off ZIP codes (or Census ZCTAs) as a starting point. They represent post offices and delivery routes, and hence the kind of transportation zones you are studying!

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.