This report examines how to enhance the automated coding of jobs and employment data within the Jobs and Employment Data Exchange (JEDx) infrastructure. Autocoding uses machine learning to automatically assign standardized codes to industry and occupation information, making it easier to analyze and compare employment data across different contexts.
The report identifies four key areas for improving occupational autocoding:
- Better training data from multiple sources to enhance coding accuracy
- Data-verified specificity to ensure coding is sufficiently detailed
- Models to support more granular coding, including skills-based classification
- Human-in-the-loop processes to validate and improve coding accuracy over time
The findings suggest that integrating improved autocoding into the JEDx infrastructure could significantly reduce manual coding effort while maintaining high data quality. The report outlines current pilot testing plans and future opportunities for leveraging JEDx data to create more precise and useful coding systems for workforce analysis and planning.
Download the full report to learn more.
