Code and data for the transition comparison

Recalculate the paper’s comparison with observed occupational transitions in Python or Stata. The download includes occupation rankings, transition totals and the values used to calculate recovery.

Download code and data · ZIP

For each starting occupation, both models recommend the same number of destinations, up to 20. This is the paper’s “equal-length” comparison. The package recalculates results from saved rankings; rebuilding the models requires the separate construction code and restricted inputs. The paper uses the three-year NLSY97/PSID horizon; five- and ten-year results are also included.

Run the analysis

Python: python replicate.py . results (requires pandas). Stata: set this folder as the working directory and run do replicate.do.

Python analysis (replicate.py)
Stata analysis (replicate.do)
Instructions and definitions (README.md)
Model-ranked destination occupations (rankings.csv)
Observed transitions by origin and destination (observed.csv)
Origin occupations and recovery denominators (origins.csv)
Reference results to check your run (expected_summary.csv)
File checksums (manifest.json)

Benchmark methods and audit

Construction choices and audit

Read the benchmark construction and audit disclosure: populations, survey pooling, crosswalks, denominators and reproduction limits. OpenAI Codex performed the 6 October 2026 audit and generated its reports; it is not independent human review.

Website layout and explanatory text were prepared with generative AI (Codex). Data sources, model methods and code.

The National Labor Exchange (NLx) Data Trust bears no responsibility for the analyses or interpretations of the data presented here. The opinions expressed herein, including any implications for policy, are those of the authors and not of the NLx Data Trust members or others.