Data Source
Job posting counts are pulled from the TheirStack API, which aggregates live US job postings across major platforms. Each data point represents the number of active postings matching a specific job title and skill combination at the time of collection.
Job Title Classification
Job titles are mapped from O*NET occupational categories — the US Department of Labor's standard framework for classifying occupations. This ensures titles like "Software Engineer" and "ML Engineer" are grouped consistently rather than relying on raw employer-written strings.
Skill Selection & Tracking
The skill list was built by identifying the most commonly requested technical skills across technology job postings, then manually curated to focus on skills that signal meaningful shifts in employer demand — particularly around AI tooling, cloud platforms, and data infrastructure. Skills are tracked as co-occurrences within a specific role type (e.g., how often "Python" appears in Data Engineer postings rather than across all jobs), which surfaces role-specific demand rather than noise from unrelated postings.
Normalization
All monthly figures are normalized to a 30-day equivalent. This corrects for the fact that February has 28 days and March has 31 — without normalization, longer months would always appear to have more postings regardless of real trend.
Update Cadence
Data is collected weekly and aggregated monthly. The dashboard defaults to monthly comparisons for stability; weekly view is available for more recent signal.
Industries
Technology and Finance are tracked separately with industry-specific title and skill configurations.
Limitations
Counts reflect job posting volume, not hires. A single open role may generate multiple postings. Figures should be read as directional signals, not absolute headcounts.