Data & Methodology

These numbers refresh from the live database — every page load reflects the most recent crawl.

Where Does the Data Come From?

We crawl public job listings daily from tech company career pages and job boards. We focus on software engineering roles across all experience levels. All data is aggregated and anonymized — we don't collect personal information from job listings.

What Counts as a “Visible” Job

Job counts are stable across the dashboard, the job board, and stack pages because every surface uses the same eligibility rules to decide which jobs to display: a job has to be active, tagged as a software-engineering role, not hidden by an admin, and have completed extraction. A job that fails any of those checks doesn't appear anywhere — including in the “Active jobs tracked” count above. This consistency is why the totals on /market, /jobs, and any /stacks/<name> page agree with each other.

Semantic Ranking on the Job Board

Beyond filtering, we add a semantic-similarity signal to ranking when you've uploaded a resume. Jobs whose content most closely matches the meaning of your resume float to the top of the list. The same signal blends with keyword search when you type a query. See Job Board for the user-side flow and Resume Matchfor the upload path. Co-occurrence and aggregate stats elsewhere don't use this signal — see Co-Occurrence & Drill-Down for the methodology there.

Update Frequency

Job data is crawled daily. After each crawl, market aggregates are refreshed: salary percentiles, tech stack demand, hiring velocity, and remote work ratios. The newsletter is sent every Wednesday with the latest weekly data.

What Gets Extracted

From each job listing, we extract and normalize:

• Tech stacks — programming languages, frameworks, and tools
• Salary data — ranges, annualized and normalized to USD
• Location — remote, hybrid, onsite, and geographic eligibility
• Experience level — junior, mid, senior, staff, principal
• Benefits — health, equity, PTO, etc.
• Culture signals — async-first, pair programming, etc.

Methodology

Salary Percentiles

Salary percentiles (p25, p50, p75, p90) are computed monthly from jobs that disclose salary information. Only annualized, US-eligible salaries above $30K are included. Hourly, weekly, and monthly rates are annualized using standard multipliers.

Hiring Velocity

Hiring velocity measures the week-over-week change in job posting volume. A positive velocity means more jobs were posted this week compared to last week.

Data Quality

Every listing goes through deduplication (by URL and title+company), validation (minimum description length, valid salary ranges), and normalization (consistent company names, standardized skill taxonomies).