DEV Community

Richard (Ryszard) Cieplechowicz
Richard (Ryszard) Cieplechowicz

Posted on Edited on

Close ZIP-area rent differences are not precise rankings

My Tampa Bay rent study publishes survey estimates for 132 ZIP-code areas: the share of renter households paying 30 percent or more, and 50 percent or more, of their household income in rent, from Census ACS 2020-2024 five-year data. Every one of those percentages carries a margin of error.

In practice, that means a small gap between two areas is not a ranking. If one ZIP area shows 40 percent of renters burdened and the next shows 42 percent, the two-point gap can sit inside the uncertainty around both estimates. (Those two numbers are an illustrative pair, not a significance test of any two real areas.) I don't treat differences like that as meaningful. I look for broad patterns instead: places where the share is clearly high across the table.

Small samples get flagged. Where a ZIP area has fewer than 300 renter units in the survey, the table marks the estimate with a caution. Blank cells mean the Census Bureau published no usable estimate for that area, not that nobody there is burdened.

A few more limits worth keeping in view. The values describe a five-year period, 2020 through 2024, not 2024 alone. ZCTAs only approximate USPS ZIP codes, and county assignment is approximate near boundaries. The percentages exclude households whose rent-to-income ratio was not computed.

A number with decimals can still be an estimate. The full table, the downloadable CSV, and the method notes are on the study page at richardcieplechowicz.com. The archived dataset is on Zenodo (DOI 10.5281/zenodo.22940611), and the source README and data are on GitHub.

Richard (Ryszard) Cieplechowicz

This article was written with AI assistance and reviewed before publication.

Top comments (0)