Occupational Structure in the U.S. Labor Market

Using large-scale LinkedIn data to map and predict occupational mobility.

  • Developed scalable Python pipelines that processed 200 M+ LinkedIn profiles, applying NLP techniques to clean text and map self-reported job titles to official SOC occupation codes.
  • Trained models on 120 M+ individual career sequences to predict when workers switch occupations, uncovering key demographic and skill correlates of mobility.