Projects

Real-time weather advisory monitoring for distributed operations teams

The Problem

Organizations with distributed physical sites — retail chains, logistics networks, field service teams — have no simple way to monitor weather threats across all their locations at once. Teams end up checking weather.gov manually, one zip code at a time. When you're responsible for hundreds of sites, that doesn't scale.

What It Does

Storm Scout consolidates real-time NOAA/NWS weather alerts into a single dashboard, giving operations leaders the situational awareness to act quickly during severe weather.

• 300 demo locations across all 50 US states and territories (swap your own via CSV import)
• Real-time NOAA data — automatic ingestion every 15 minutes, 96 alert types classified across 5 severity levels
• 6 dashboard pages — Overview, Active Advisories, Offices Impacted, Notices, Filter Settings, Sources
• Smart filtering with 5 presets and persistent user preferences
• Interactive map with severity-aware marker clustering
• Office detail views with full NOAA alert narratives and nearest weather station observations

The Development Story

Storm Scout was built by an operations leader — not a software engineer — using AI-assisted development. 275+ GitHub Issues, 300+ commits, approximately one month. Every commit carries a Co-Authored-By trailer identifying the AI tool used (Claude, GitHub Copilot, Gemini, Warp.dev). I directed all architecture, security, and UX decisions; AI tools generated implementations against well-scoped issues with explicit acceptance criteria.

Tech Stack

Engineering Highlights

• Performance: in-memory caching, gzip compression (~85% payload reduction), client-side TTL cache, pagination, connection pooling
• Security: CSP headers, SRI hashes, tagged template XSS prevention, timing-safe API key auth, input validation, rate limiting
• Reliability: circuit breaker for NOAA API, pre-deploy smoke tests, CI pipeline via GitHub Actions