How to Compute Characteristics at the Zip Code Level

Zip codes are not census geography so the Census Bureau creates something called ZipCode Tabulation Areas built from lower level census geography (census blocks). Users can do this as well, although we recommend using census tracts as the lower level of geography. Very little tabular data are available at the block level with the ACS.

There is a census tract to zip code cross-walk available from Mable/Geocorr out of the Missouri State data center.

This tool provides correspondence between census geographies from 1990, 2000 and 2010.

If you are interested in producing zip code characteristics, follow theses step:

Step 1: Create a weights file from the appropriate Mable GeoCorr site (most likely 2010). Select your state(s) of interest; choose ZCTA as SOURCE and Census Tract as TARGET.

The weights you get for a single zip code will sum to 1. As an example, see the weights for zip code 01001 in Massachusetts (column I):

Step 2: Pull off the characteristics you are interested in. In our example, we used median and mean household income from the 2006-2010 ACS. [Zip code data are now available from the 2007-2011 release.] Our source was Social Explorer, Tables T57 and T59 from the 2006-2010 file. A publicly available source for the same information is Tables S1903 and S1902 from American FactFinder.

Step 3: Merge the characteristics file with the Weights file (by census tract).

Step 4: Weight the characteristic(s) by the zip2tract weight and sum the product.

This is illustrated in the example below. Column G is the product of the zip2tract weight (column D) and Mean Household Income (column F). The sum of the product is $67,480. That is the mean household income for zip code 48111.