Landscape

A living map of the creative industry — starting with Israel. Studios, production companies, creative-tech startups, agencies and the brands that buy creative work, in one searchable, source-backed data layer that keeps itself up to date.

• Use it: the claude.ai app (private link) — Explore brands & collaborators ranked by fit, a weekly Signals feed, and your private Pipeline. Or GitHub Pages / python3 -m http.server -d site (pipeline saved on that device only).
• Spreadsheet: data/exports/landscape.csv
• Coverage: STATS.md · Plan: PLAN.md

How it works

research (Claude, weekly or on demand)
   └─> research/staging/*.json ──ingest──> data/orgs/<id>.json ──build──> site/ + exports
                                  (dedupe, merge, validate)              (explorer, CSV, stats)

Each research run opens a pull request, so new findings are reviewed before they land.

Run research on demand

• In a Claude Code session on this repo: "run research on il-vfx-post" or "find design studios in Haifa".
• Or on GitHub: Actions → Research → Run workflow (needs a CLAUDE_CODE_OAUTH_TOKEN or ANTHROPIC_API_KEY repo secret).

Segments and their order live in research/queue.md.

Commands

pip install jsonschema
python3 scripts/landscape.py validate          # check all records
python3 scripts/landscape.py ingest FILE.json  # merge new/updated records
python3 scripts/landscape.py signals FILE.json # merge new signals
python3 scripts/landscape.py logos             # fetch logos for new orgs (pip install pillow)
python3 scripts/landscape.py build             # rebuild app data, CSV, STATS.md
python3 scripts/landscape.py stale --days 90   # what needs re-verifying
python3 -m http.server -d site 8000            # browse locally at localhost:8000

Publish the app on GitHub Pages

Settings → Pages → Source: GitHub Actions. Every merge to main redeploys it.

Record format

One file per organization in data/orgs/. Required: id, name, kind, categories, description, location, sources, confidence, last_verified. Optional: Hebrew name, founded, size, website & social links, services, notable clients & work, key people (public leadership only), funding, client profile (industry, creative needs, agencies), relations (client / parent / acquired_by / partner…), tags. Full schema: schema/org.schema.json.