Public goods

A few bits of code I’ve written over the years for fairly standard tasks that were annoying and time-consuming. I am confident in the accuracy of the code for my specific purposes, but no warranty whatsoever is offered for use elsewhere.

Data cleaning

Cleaning the Consumer Expenditure Survey

Stata code that cleans the CEX following Aguiar and Bils (2015). Handles top-coding, multi-quarter aggregation across the interview and diary surveys, and the 2013 redesign. Outputs a household-quarter panel with consistent expenditure categories from 1996 onward.

Cleaning the PSID

R code that cleans the PSID following Guvenen and Smith (2014). Constructs annual measures of earnings, wealth, and consumption with consistent definitions across the biennial waves post-1997.

Geography

Linking historical UK geographies

A Bash + Python pipeline that uses the ONS Geography Portal to crosswalk administrative geographies across boundary changes — parishes, wards, and LSOAs — with population-weighted apportionment for variables that don’t aggregate cleanly.