Data Sources
All data in this visualization comes from official, freely accessible sources. Here you can see exactly where the numbers come from — including direct download links.
Primary Data Sources (directly used)
These three datasets are automatically downloaded and fed into the visualization. All download links lead directly to the official files.
Official EU employment statistics: how many people work in each industry sector in Austria.
What we use from this: Total employment per NACE Rev.2 sector (e.g. Manufacturing, IT, Health, etc.)
Official Austrian earnings survey: what people earn per hour in each industry, broken down by sector.
What we use from this: Gross hourly median pay by sector (full-time workers, all genders). Statistik Austria publishes this survey under ÖNACE 2008 / NACE Rev.2 codes (see sector codes CSV, e.g. C-ONVE10-0). We map rows to section letters A–S using the same structure as ÖNACE 2025 (identical at section level).
Official Austrian earnings survey: what people in specific occupations earn (e.g. nurses, software developers, teachers).
What we use from this: Gross hourly median pay by ISCO-08 occupation group (full-time workers, all genders)
How We Process the Data
- Employment: Sector totals taken directly from Eurostat (exact). Split into individual occupations = estimated proportions within sectors.
- ÖNACE / sector labels: We present sectors using ÖNACE 2025 (current since 1 Jan 2025) for A–S labels. VSE 2022 earnings were collected under ÖNACE 2008; at section level the structure matches ÖNACE 2025—differences are mainly at detailed (4–5 digit) subclasses.
- Salaries: Median gross hourly pay from VSE 2022 × 2,080 hrs/year × 1.17 (13th + 14th salary) = annual gross salary.
- AI Exposure: Integer 0–10 plus English rationale per occupation group, defined in scripts/generate-occupations.ts (OCCUPATION_DEFS). Rubric matches Karpathy’s concept (cognitive/digital task content; not empirical measurement like Felten et al.). No LLM API call at build time—values are curated and written to data.ts when you run the generator. Optional: npm run score:exposure-llm (OPENROUTER_API_KEY) writes scripts/llm-exposure-overrides.json; the generator merges those for exposure + rationale only.
- Job Outlook: Integer –10…+10 and short label (outlookDesc) in OCCUPATION_DEFS; qualitative demand signal per aggregated group, sector-informed—not identical to AMS forecasts, but in the same policy discourse (structural change, WIFO/AMS).
Additional Reference Sources
Monthly labor market reports: registered unemployment, employment figures, vacancies
Labour Force Survey, register-based statistics, job vacancy statistics
Employment statistics by chamber system, industry data, ÖNACE classification
Open government data portal with CSV/JSON/API access
Federal Ministry labor market data and reports
Additional Sources
- • ÖNACE 2025 (current standard since 1 Jan 2025): WKO PDF | Statistik Austria EN
Note: VSE 2022 earnings files use ÖNACE 2008 codes; sections A–S are structurally the same.
- • Eurostat Labour Market: EURES Austria
- • IMF World Economic Outlook 2025 (GDP per capita, PPP comparisons)
- • OECD Employment Outlook (working hours, labor force participation)
- • Statistik Austria Allgemeiner Einkommensbericht 2024 (salary data)
- • WIFO (Wirtschaftsforschungsinstitut) labor market projections 2025-2030