Workforce Patterns and Income Disparities in a Growing US City

Open data and code for Berg et al. "Workforce Patterns and Income Disparities in a Growing US City"

Input data are from the U.S. Census Bureau's Public Use Microdata Sample (PUMS) and are in the public domain. Derivative data and code is published under a Creative Commons Attribution 4.0 International License. For attribution, please cite the article:

Aleksander K Berg, Anthony Howell, Sharmistha Bagchi-Sen, Workforce patterns and income disparities in a growing US city, Journal of Economic Geography, 2026;, lbaf071, https://doi.org/10.1093/jeg/lbaf071

Files and what they do:

1. Code - contains all data cleaning, modeling, and visualization R code (run in this order). Data processed in R version 4.5.1.
   1. usa_io_2013.r - USA 2013 - contains code for counting total workers in each IO, data cleaning for shift share
   2. usa_io_2019.r - USA 2019 - contains code for counting total workers in each IO, data cleaning for shift share
   3.  atl_io_2013.r - Atlanta 2013 - contains code for counting total workers in each IO, data cleaning for shift share
   4. atl_io_2019.r - Atlanta 2019 - contains code for counting total workers in each IO, LQ calculations for 2019, and data cleaning for shift share
   5. atl_shiftshare_lq.r - contains code for the classification of IO catagories based on shift share and lq
   6. atl_summarystats.r - contains code to create summary statistics tables
   7. summarystats_tests.r - contains code performing Chi-Square, Wilcoxon, and Welch two-sample tests for the Appendix
   8. atl_models.r - contains all model code
2. Data - contains all input, working, and output data from the analysis
   1. inputs
      1. usa_pums_2013.csv - raw 1-year 2013 ACS PUMS files for the US, downloaded via the U.S. Census Bureaus FTP archive
      2. usa_pums_2019.csv - raw 1-year 2019 ACS PUMS files for the US, downloaded via the U.S. Census Bureaus FTP archive
   2. working - contains derivative files created during the analysis. See codebooks for details about generation
   3. outputs - contains output tables created automatically during the analysis. Some tables in the manuscript were created manually from tables in the "working" folder