CHAOS
Converting occupational classifications into continuous income estimates — formalizing and scaling what economic historians have done informally for decades.
Status: Active — early results, presenting at conferences
Economic historians routinely need to convert occupational categories into income or status scores to study inequality, mobility, and structural change. Current approaches are inconsistent and hard to replicate. CHAOS (Classification Hierarchy to Approximate Occupational Scores) formalizes this into a principled, replicable ML pipeline.
CHAOS takes HISCO occupational codes — the output of OccCANINE — and maps them to continuous income estimates calibrated against historical wage and tax data. The contribution is both methodological (a rigorous estimator where ad hoc scores have been used) and infrastructural (it completes the pipeline from raw occupational text to analyzable income data).
CHAOS is a direct input to the Copenhagen Income project and Parish Tax project.
With co-authors.