Occupational models from 42 million unstructured job postings

Structuring jobs into occupations is the first step for analysis tasks in many fields of research, including economics and public health, as well as for practical applications like matching job seekers to available jobs. We present a data resource, derived with natural language processing techniques from over 42 million unstructured job postings in the National Labor Exchange, that empirically models the associations between occupation codes (estimated initially by the Standardized Occupation Coding for Computer-assisted Epidemiological Research method), skill keywords, job titles, and full-text job descriptions in the United States during the years 2019 and 2021. We model the probability that a job title is associated with an occupation code and that a job description is associated with skill keywords and occupation codes. Our models are openly available in the sockit python package, which can assign occupation codes to job titles, parse skills from and assign occupation codes to job postings and resumes, and estimate occupational similarity among job postings, resumes, and occupation codes. •Job titles follow conventions and can be normalized to reduce variation•We curated a list of 775 skills and estimated their probability by occupation•Skill probabilities are consistent with the hierarchy of occupational codes•Job posting counts are an imperfect proxy for job openings reported in survey data Online job postings offer an abundant and detailed view of occupations and skills in the labor market. However, the variation in how employers refer to and describe job openings makes it difficult to use unstructured job postings for analysis and research. Occupations have long been standardized in the United States into a hierarchy of 867 codes through the Standard Occupational Classification system. By categorizing a sample of 42 M United States job postings into these standardized occupational codes and extracting the skills in each posting, we constructed an open dataset with empirical probabilities for associations among occupational codes, job titles, and skills. We bundled these data in a software tool, called sockit, that can analyze new job titles, job descriptions, or resumes. We present an open data resource and software tool for understanding the associations between occupations, job titles, and skills in the United States labor market. These associations are used in several fields of research, including economics and public health, as well as for prac