Historical U.S. County Population Data: 1900 to 2010
Update
This dataset has gotten a fair bit of attention from fellow social scientists. Some have even asked for a way to cite it! I’ve updated the data (to now include 2020 decennial populations) and have uploaded it to the Harvard Dataverse. Anyone who’d like a copy of my data can access it here. Please use the following citation:
Van Leuven, Andrew J, 2024, “Historical U.S. County Population Data: 1900 to 2020”, https://doi.org/10.7910/DVN/WLS5GF, Harvard Dataverse, V1.
These data—1900 to 1990 county populations—come from the U.S. Census (original link is now dead) but have not historically been available in tabular form. I wrote a custom R script1 which converted the individual state-level .txt files from this Census webpage into usable .csv files. The county population data for 2000 and 2010 were collected using Kyle Walker’s excellent tidycensus package.
I hope this project is a helpful time-saver for anyone in the future who needs this dataset.
Data
Here is a direct download of the data.
Here is a link to the data in GitHub (as a searchable online table).
Here is my R code which I wrote for this project. The custom function I created did not work for four states (AK, CO, FL, and SD) so I processed those .txt files by hand, using Excel. Still, 47 out of 51 isn’t bad! If anyone has an idea why my script couldn’t handle those states, please reach out to me. ↩︎
