Data Verification

Every figure on this site is checked against original data from Eurostat and Statistics Austria. 99 hypothesis-driven tests verify employment, earnings, occupation structure, AI exposure, and sector totals — automatically on every build.

Source-validated

13 tests compare generated data directly against cached originals from Eurostat (lfsa_egai2d, nama_10_a64_e) and the Structure of Earnings Survey 2022 from Statistics Austria (OGD_veste403). These verify: ISCO employment sums, NACE sector totals, VSE pay matching, ranking correlation, and data freshness.

Internal consistency

86 tests check internal consistency of generated data: field completeness, ISCO-08 structural rules, employment distribution (HHI, skewness), pay plausibility (Gini, P90/P10), AI exposure distribution, outlook bias, and cross-dimensional validation.

Scientific methodology

Each test follows a scientific format: a thesis states the expected property of the data (e.g. "Professionals have higher AI exposure than craft workers"). An antithesis describes what failure would mean for data quality. The test then produces a verdict with concrete evidence. Click any individual test to inspect its thesis, antithesis, and evidence.

Data Verification

99/99 passed

99 hypothesis-driven tests verify employment totals, earnings plausibility, ISCO consistency, AI exposure scoring, and labor-market structure.

Data Completeness 17/17ISCO-08 Structure 8/8Employment Distribution 9/9Sector Totals 13/13Earnings & Pay 14/14AI Exposure 10/10Outlook & Growth 9/9Cross-Validation 6/6Source Data (Eurostat/VSE) 13/13

Source data cached in src/lib/real-data.ts (last fetched: 28 March 2026). To refresh: npx tsx scripts/fetch-real-data.ts && npx tsx scripts/generate-occupations.ts

All source data is freely downloadable as Open Government Data. The pipeline is deterministic — no LLM is called at build time.