Document resource
Objectives The goal of the International Partnership on Automatic Job Coding (IPAJC) is an interdisciplinary partnership aiming to enable and promote efficient collection, processing, and accessibility of job information through improvement of automatic coding tools.Methods and Results Partners within IPAJC create and apply various automatic and semi-automatic coding tool using different methods. Examples include fuzzy matching and Boolean operators (e.g. CASCOT, University of Warwick), machine learning (ML) neural network (e.g. NIOCCS, US NIOSH), ML boosted decision trees (e.g. OccuCoDE, Ludwig Maximilian University, Munich; OPERAS, Utrecht University), ML ensemble classifiers (e.g. SOCcer ,US NCI and AUTONOC, University of New Brunswick/Dalhousie University), and large language models (e.g. TNO-AOC, TNO). Depending on the specific tool, free text job descriptions in different languages are converted into standardised occupation codes such as various versions of international ISCO, Canadian NOC, German KldB, UK SOC, and US SOC. Although most tools thus far focused on coding to standardised occupations, some tools for coding to standardised industries are also available (e.g. CASCOT; NIOCCS; SOCcer CLIPS).Conclusion The IPAJC continually seeks partners and job description data for development and validation of current and future automatic job coding tools. Our expertise on survey methods, occupation data collection, and occupation coding is improving job data collection and exposure assessment at scale in occupational health research.