Add comprehensive data collection documentation

[Copilot]

Users and contributors need visibility into the app's data sourcing, collection methods, and architecture to understand system behavior, debug issues, and add new sources.

Changes

Created /docs/DATA_COLLECTION.md (35KB) documenting the complete data pipeline:

Data Sources (150+)

• Conflict/Geopolitical: ACLED, UCDP, GDELT, UNHCR
• Real-time Tracking: AIS Stream (vessels), OpenSky (aircraft), Wingbits enrichment
• Natural Disasters: GDACS, NASA EONET/FIRMS, USGS earthquakes, NWS alerts
• Markets/Economics: Finnhub, Yahoo Finance, CoinGecko, Polymarket, FRED, EIA
• Infrastructure: Cloudflare Radar outages, NGA cable warnings
• Tech/Research: ArXiv, GitHub trending, Hacker News
• Cybersecurity: Threat feed aggregation, CISA advisories
• News: 100+ RSS feeds (wire services, government, think tanks, regional)

Architecture & Methods

• Three-tier infrastructure (Vercel edge → Railway relay → Browser services)
• Collection methods: REST APIs, RSS parsing, WebSocket streaming, GraphQL
• Caching strategy: Upstash Redis (cross-user) → localStorage (per-user) → in-memory
• Refresh intervals: 2 minutes (markets) to 1 hour (research papers)

Operational Details

• API authentication requirements and key management
• Circuit breakers and multi-source fallbacks
• Rate limiting and error handling
• Data processing pipeline (ingestion → transformation → classification → aggregation)
• Performance optimizations (lazy loading, progressive disclosure, WebWorker offloading)

Developer Resources

• Troubleshooting guide for common data issues
• Guidelines for contributing new data sources
• License and attribution requirements per source

Integration

• Added "Data Sources" link to README.md header navigation
• Added prominent callout in DOCUMENTATION.md pointing to new doc

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.