JobsCY — Cyprus Job Market Visualiser

An interactive treemap visualisation of the Cyprus labour market, with AI exposure scoring for occupations. Built on data from EU/Cyprus sources including HRDA/AnAD, Eurostat, and EURES Cyprus.

Created and maintained by Alexandros Zenonos, adapted from karpathy/jobs for the Cyprus labour market using EU data sources and international classifications (ISCO-08, NACE, ISCED).

What's here

An interactive treemap of 39 ISCO-08 occupation groups covering the Cyprus labour market. Each rectangle's area is proportional to employment and colour shows the selected metric — toggle between growth outlook, median pay (EUR), education level, and AI exposure.

Data is sourced from the Eurostat SDMX REST API (employment by ISCO-08 2-digit, earnings by ISCO-08 1-digit, filtered to geo=CY) and enriched with HRDA 2022-2032 occupation forecasts.

Data sources

Source	Dataset	What it provides	Link
Eurostat	LFSA_EGAI2D	Employment by ISCO-08 2-digit, filtered to Cyprus (2024)	SDMX API docs
Eurostat	earn_ses_hourly	Mean gross hourly earnings by ISCO-08 1-digit, Cyprus (SES 2022)	Dataset explorer
HRDA/AnAD	—	309 occupation forecasts (2022-2032), expansion + replacement demand	anad.org.cy
EURES	—	Shortage/surplus occupations, vacancy statistics by ESCO occupation	eures.europa.eu
CEDEFOP	—	Skills forecasts to 2035 by sector and occupation group	cedefop.europa.eu
CYSTAT	—	Aggregate employment, unemployment, labour costs	cystat.gov.cy

Earnings methodology

Salary figures are derived from the Eurostat Structure of Earnings Survey (earn_ses_hourly), the official EU-wide enterprise survey on wages:

• Granularity: SES provides earnings at ISCO 1-digit (major group) level only — all 2-digit occupations under the same group share the same figure (e.g. all Professionals = €36,878/year)
• Conversion: Hourly earnings × 2,080 hours (40 h/week × 52 weeks) = annual pay
• Coverage: Enterprises with 10+ employees — excludes self-employed, micro-firms, and the informal economy
• Survey year: 2022 (SES is published every 4 years; next wave: 2026)
• Metric: Mean gross hourly earnings in EUR (MEAN_E_EUR)

These are statistical averages, not market salaries. Actual compensation for specific roles in Cyprus may differ significantly due to sector, firm size, seniority, and private-sector premiums not captured by SES.

Architecture

Cyprus pipeline:
  ISCO-08 data ──► generate_cy_occupations.py ──► occupations_cy.json
  Eurostat API ──► make_cy_csv.py ─────────────► occupations_cy.csv
                                                        │
  LLM (OpenRouter) ──► score.py ──► scores.json ◄──────┘
                                         │
                     build_site_data.py ──┘──► site/data.json ──► site/index.html

LLM-powered colouring

The repo includes a pipeline for scoring occupations using LLMs via OpenRouter. You write a prompt, the LLM scores each occupation, and the treemap colours accordingly. The "Digital AI Exposure" layer estimates how much current AI will reshape each occupation within the Cyprus/EU labour market context (considering tourism, shipping, financial services, EU Digital Decade targets).

Fork score.py to write your own scoring criteria — e.g. green economy relevance, remote work potential, EU Digital Decade alignment.

What "AI Exposure" is NOT:

• It does not predict that a job will disappear
• It does not account for demand elasticity, regulatory barriers, or social preferences
• The scores are LLM estimates, not rigorous predictions

Key files

File	Description
eurostat.py	Eurostat SDMX 2.1 REST API client (employment + earnings data)
generate_cy_occupations.py	Generate occupations_cy.json from ISCO-08 classification
make_cy_csv.py	Build occupations_cy.csv from Eurostat data (EUR, ISCO-08)
score.py	LLM-based AI exposure scoring via OpenRouter (ISCO-08 / Cyprus context)
build_site_data.py	Merge CSV + scores → site/data.json
make_prompt.py	Generate single-file LLM prompt from all data
site/index.html	Interactive treemap visualisation (EUR, ISCO-08, Cyprus)
occupations_cy.json	Master list of 39 ISCO-08 occupation groups
occupations_cy.csv	Summary stats: pay (EUR), employment, education, ISCO codes
scores.json	AI exposure scores (0-10) with rationales

Setup

uv sync                  # install dependencies
uv sync --extra dev      # includes pytest, ruff for development

Requires an OpenRouter API key in .env (for LLM scoring only):

OPENROUTER_API_KEY=your_key_here

Usage

# Generate occupation list from ISCO-08
uv run python generate_cy_occupations.py

# Fetch Eurostat data and build CSV
uv run python make_cy_csv.py

# Score AI exposure (uses OpenRouter API)
uv run python score.py

# Build website data
uv run python build_site_data.py

# Generate LLM analysis prompt
uv run python make_prompt.py

# Serve the site locally
cd site && python -m http.server 8000

Development

uv run pytest -v           # run tests (110+ tests)
uv run ruff check .        # lint
uv run ruff format .       # auto-format

CI runs automatically on push via GitHub Actions (lint + test + site validation). See CONTRIBUTING.md for development guidelines.

Deployment

The site is deployed as a static site on Vercel. Configuration is in vercel.json:

• Output directory: site/
• No build step — the site is pre-built and committed to the repository
• Security headers: HSTS, CSP, X-Frame-Options, X-Content-Type-Options, Referrer-Policy, Permissions-Policy
• Caching: data.json is cached for 1 hour with 24-hour stale-while-revalidate

Currency and classifications

• Currency: EUR (€) — all monetary values
• Occupations: ISCO-08 (International Standard Classification of Occupations)
• Sectors: NACE Rev. 2 (Statistical Classification of Economic Activities)
• Education: ISCED 2011 (International Standard Classification of Education)
• Data period: Eurostat latest available year, HRDA 2022-2032 forecasts

Acknowledgements

This project adapts the visualisation approach from karpathy/jobs (a US BLS occupational treemap) for the Cyprus labour market using EU data sources and international classifications.