feat: data map (ontology), current-issue analysis guide, plain-language polish

[jsshin2022]

What's added

• Data map (ontology) — catalog/datasets.yaml: 35 public datasets. Each has:
  • its role in an analysis: treatment (who was treated, when), outcome (what changed), covariate (other factors);
  • space × time grain, access method and approval type, license, and the date it was checked.
  • 32 of the 35 were checked on the original data.go.kr or Seoul Open Data page. The 3 KOSIS tables are verified by title only and are marked as such.
• Current-issue guide — catalog/issues.yaml: 6 policies debated in 2025–26:
  • 10·15 and 3·24 land-transaction permits, the 6·27 mortgage cap, the consumption coupon, the high-oil support payment, and K-Pass "모두의 카드".
  • Each issue has an official start-date source, treated and comparison groups, outcomes, the recommended design with its support status, collectable datasets, and pitfalls.
• Two new topics: consumer-support and public-transit, so every issue belongs to a topic. This keeps the topic-first rule.
• core/discovery/ontology.py: schema, loaders, and a cross-reference check run in CI (no dangling topic or dataset ids).
• Site
  • New pages: 지금 이슈 (5-step estimation guide, plain-language design explainers, a guide per policy) and 데이터 지도 (relation diagram, search and filters by role, topic, grain and access, a data.go.kr search box, portal list, glossary).
  • Home, topic and architecture pages rewritten in plain Korean.
  • CSS moved to site/style.css.
• Docs: README, catalog/README.md and the group guide rewritten for mentees.

Verification

• pytest: 49 passed; ruff: clean; the site builds with no JS errors; all pages fit a 390px width.

🤖 Generated with Claude Code

https://claude.ai/code/session_01EMnmZPBD5PQJW1UQAARN7q

---

Generated by Claude Code