Zip Code Median Income Data

I'm having a heck of a time downloading median income for zip codes on the ACS 5-year surveys. I know I'm doing something wrong but can't figure it out. I choose the survey, go to geographies, click five digit zip code tabluation area and no zip codes show up. Forgive my impertinence, but can someone help? Thank you.

Parents Reply Children
  • What do you mean? ZCTAs don't have names.

  • I've done that by matching on USPS data -- not impossible -- they tend to have multiple names, let me see if I can reconstruct what I did

  • I happened to answer this very question on the Census Slack today:

    The Census Bureau doesn’t assign place names to ZIP Code Tabulation Areas (which are based on, but not the same as ZIP Codes), and there’s no rule that says that a ZCTA must be within a single state. Of course, people still do what you’re asking all the time — but you can’t use Census data to do it.This writeup, describing a process we used when we needed to relate ZIP Codes to ZCTAs, also gets a bit more into the difference between them and links to other resources. As part of that process, we used GeoNames, a free source of a ZIP Code dataset which does assign a name to each ZIP Code.In short, while not all ZIP Codes have a matching ZCTA, each ZCTA should match to a ZIP Code in the GeoNames dataset which will have a name; you will just have to load that data separately from API calls.
  • Just keep in mind, ZCTAs are only updated every 10 years, and ZIP codes are constantly changing based on changes to post office locations, mail routes, etc. Since it's now 2021, and the most recent ZCTAs are from 2010, ZCTAs are currently as out-of-date as they get in this cycle. So if your ZIP code source is current, the geographies represented by matching codes could be significantly different.

    In 2016, I did an analysis matching ZCTAs with ZIP codes, and found that 70% of ZCTAs shared at least 80% of their area with their corresponding ZIP codes. Not bad, but 11% of ZCTAs shared less than half their area with their corresponding ZIP code. And since that was five years ago, the numbers are surely worse now. You can read the blog post (mostly meant for beginners) here:

    Just for fun, these were some examples of ZCTAs that didn't really match their ZIP codes:

  • This is all true! Especially when a zip is at the margins of civilization (in the Miami area you'd have vast swaths of Everglades that would suddenly turn into subdivisions) but for my purposes a few years ago it worked well, maybe partly because I was comparing home price index values from decades ago so they tended to be suburban areas that hadn't changed much in that time -- I was looking at zip code level index, using ACS to gauge black homeownership, and the finding on inequality stood up well when I went to some of those zip codes and looked at their histories

  • To me, this maybe sounds like a job for the Univ. of Missouri's Geographic Correspondence Engine -- it's essentially the crosswalk to end all crosswalks -- it'll provide you a list of all ZCTAs by Census Place (or vice versa) for a given area in .CSV format.  Here is a link for HTML output of results from a quick run I just did for Washington.  

    More often than not, there will be multiple Census Places per ZCTA, but the results do include populations for the places which would allow attribution to the most populous place as a best possible outcome.  


  • And, if I'd read the entire thread, I would have seen that @Glenn_Rice made this suggestion 4 months ago; so consider this my concurrence.