The Replaceability Index Dataset: Access, Fields, and Limits
By Shawn Geeman · 2 min read · Published · Updated
The public dataset makes the published occupation values inspectable. It combines a BLS-based occupational dataset with Index exposure scores. The exposure values are our estimates; they are not official BLS or OECD forecasts.
Download the data
The CSV is available at /data/replaceability-index.csv. The JSON export is at /data/replaceability-index.json. Both use the same public occupation set, deduplicated by title.
The export date records when the file was generated. It does not prove that the underlying wages, projections, or exposure estimates were refreshed on that date. Preserve a copy of the file used for any published analysis.
The nine CSV fields
soc identifies the occupational classification; title is the display name; replaceability_score is the 0 to 100 relative exposure value; tier is the corresponding product band.
median_wage_usd and hourly_wage_usd describe wages in the source dataset. projected_growth_pct is the source employment-growth projection. employment_volume is the employment count. description provides occupational context. Check source periods before using these as current market values.
What the export does not contain
The CSV does not provide individual worker records, employer adoption plans, actual layoffs attributed to AI, or a measured time allocation for every task. It also does not reproduce model-generated personal reports.
The exposure scores are curated estimates. Inspecting the output is useful, but full independent reproduction would also require the assumptions, mappings, and calibration records behind each estimate. A downloadable table alone is not evidence that a model is validated.
Checks before analysis
Check the row count, duplicates, missing values, field types, and score range. Decide whether your summary is weighted by occupation or employment. Keep wage projections separate from any assumptions about future AI adoption.
Correlation with another exposure study measures agreement between models under the chosen matching procedure. It does not establish accuracy against future employment outcomes. Read the comparison cautions.
Citation and reuse
The public export is offered under CC-BY-4.0. Attribute the Replaceability Index, link to the source, identify the file version or access date, and explain transformations you make. Consult the license text for its terms.
Cite original BLS, O*NET, OECD, or academic sources separately when your claims rely on their research. We do not claim a separately published MIT-licensed scoring repository or a guaranteed update schedule.
Frequently asked questions
Is the exposure score an official government statistic?
No. Wage and employment context comes from the source occupational dataset; the exposure score is an Index estimate.
How often is the dataset updated?
No fixed refresh interval is promised. Check source dates and retain the downloaded version. A newly generated export is not necessarily a newly calibrated dataset.
Is the personal assessment reproducible from the CSV?
No. The personal assessment can involve model generation and user-specific task descriptions, which are not part of the public export.
Sources and methodology
The score is a relative index of task exposure from 0 to 100. It is not a percentage chance of losing a job or a forecast of layoffs. Individual tasks, employer choices, adoption, and regulation all matter.
The Index uses stored BLS-based occupation data. O*NET, OECD capability indicators, and task-exposure research are contextual references. The derivation of the stored occupation scores is not reproducibly documented in the published dataset.
The source snapshot does not record the wage-survey year or employment-projection window. Check the original BLS release before treating these figures as current. Editorial changes and export dates do not establish data freshness.
This essay reflects the evidence available when written; future outcomes remain uncertain.
Discover your free score · Explore the $5 Pro Report · All essays