Review of Burning Glass Job-ad Data
Report
Posting and searching online job-ads is ubiquitous in the U.S. labor market for both employers and job seekers. Job seekers who search online job-ads are much more likely to find work and find work faster than those who do not (Faberman & Kudlyak, 2016). In addition, the job-ads posted by employers on corporate websites and online job boards provide a source of opportunity data that has opened up new avenues for research. This technical document evaluates the feasibility of using the real-time labor market information (LMI) collected by Burning Glass Technologies to supplement survey and administrative data collected by federal and state governments. In contrast to designed and administrative data which have varying lag times between collection and dissemination and are often aggregated over broad occupation groups, the real-time job ads collected by Burning Glass Technologies are made available within a day of a job-ad being posted and provide information at a granular level that links employer skill set requirements to detailed occupations in the O*NET-SOC taxonomy1. In this document, the data are evaluated for use in identifying the skill sets necessary for a job in the skilled technical workforce, future work will evaluate the Burning Glass Technologies resume data in defining pathways to skilled technical workforce jobs.
This document reviews the use of the Burning Glass Technologies job-ads in academic research highlighting issues with the data and if the researchers made any attempt to validate the Burning Glass Technologies job-ads data, the validation data sources they used, and their results. It provides the results of profiling and exploratory data analyses, for both the Virginia Burning Glass Technologies job-ads data and the Virginia Open Data/Open Jobs Data2 which is compared to the Burning Glass Technologies job-ads data. The document concludes with recommendations regarding fitness-for-use.
Burning Glass, job postings, academic research, exploratory data analysis (EDA), O*NET-SOC
English
University of Virginia
March 10, 2021