Has anyone used the Postal change of address data? Interesting compare I would think.

The files are at;


2022 is 278k records.

The data is certainly interesting and has some flaws, but is full data (no sample).  Has anyone done some analysis \ published \ dashboard etc.?



  • Careful — a lot of data suppression as you’ll see if you try adding up nationally or by state — more outmoves than inmoves meaning a lot of inmoves are unaccounted for. Cells less than 10 suppressed is the apparent reason, you can make some dent by estimating family size but that creates its own problems and net moves are still impossibly negative in many fast growing areas most notably Maricopa County AZ. Another theory is that college students don’t fill out the forms when they move in (freshmen are teenagers and don’t get much mail) but do fill in forms when they move out (they have job applications out and need forwarded mail more at that point in their lives). 
    Having said all that I found them useful in some ways — you could read a sudden spike in outmoves from dense urban zips in March 2020 and that told a story that looks was later backed up by Census county estimates. I felt more confident about this result going in because it mirrored some proprietary data from surveys of home buyers reported by researchers in this story.


  • when dealing with data, it's common to perform exploratory data analysis (EDA) to better understand its properties, look for patterns, and identify potential flaws. This could involve visualizing the data using graphs and charts, calculating summary statistics, and checking for missing or incorrect values. After the EDA, researchers or analysts might conduct more formal statistical tests or build predictive models to gain insights from the data.