Job Ad Analysis Toolkit (JAAT) Evidence Dashboard

Job postings and new open-source language models let us test old maps of work and build new maps. This dashboard summarizes public aggregate evidence from the NLx corpus and USAJOBS job postings. The main findings are A. A New Map of Work, B. Analyzing the Mismatch, C. Finding Emerging Skills, and D. Dynamic Maps, with corpus-specific drilldowns under NLx corpus and USAJOBS.

Use the profiler pages to inspect occupations, industries, agencies, tasks, skills and concepts. For the current paper’s figures and benchmarks, visit the manuscript companion.

NLx data disclaimer. The dashboard uses aggregate evidence derived from the National Labor Exchange (NLx) corpus; raw NLx posting text is not public. The National Labor Exchange (NLx) Data Trust bears no responsibility for the analyses or interpretations of the data presented here. For more information, see NLx Research Hub.
New visual story. Start with the full-screen narrative view of task diffusion and adaptive taxonomies: open the story page.
Try the tools and inspect the project. JAAT.app · Working paper · GitHub · Hugging Face

Corpus At A Glance

NLx approx. source postings

~162M

NLx corpus monthly active jobs

416.5M

NLx corpus O*NET tasks extracted

813.3M

USAJOBS raw postings

3,089,366

USAJOBS monthly active jobs

4.1M

USAJOBS O*NET tasks extracted

30.6M

The NLx source-posting count is an approximate distinct-posting scale before monthly expansion. Dashboard files contain aggregate evidence only; raw NLx posting text is not public. Monthly active jobs are posting-month rows, so one posting can contribute to more than one month.

Finding Map

Drill-Down Profiles

Methods and Audit