SupplyWatch

SupplyWatch

Inspiration

Global shipping is one of the most fragile critical systems in the world. A single disruption at a port can trigger delays across entire supply chains. We were inspired by the gap between data visibility and decision speed — teams often have data, but not a fast, actionable way to reroute operations. So we built SupplyWatch to answer one core question:
Can we detect route risk early, explain it clearly, and recommend a safer route before the disruption escalates?

What it does

SupplyWatch is an AI-powered route intelligence platform for maritime logistics. It lets operators:

View a live global network of ports and routes
Monitor route-level risk in real time
See whether risk changed due to AI/news signals or stayed stable
Understand risk drivers through explainable route intelligence
Highlight safer route alternatives between the same origin and destination ports In short: Detect risk. Explain change. Recommend action.

How we built it

We built SupplyWatch as a full-stack application:

Frontend: React + TypeScript + Vite + Three.js for an interactive globe UI
Backend: Node.js + Express REST API
Database: Supabase (PostgreSQL) for routes, ports, and risk snapshots
Risk pipeline: News ingestion + scoring logic + snapshot generation
Optimization layer: AI-assisted safer-route selection with explainable outputs
UX features: Route filtering, pagination, alerts, trend sparklines, route intelligence panel, port focus interactions, and highlighted detours

Challenges we ran into

Geospatial normalization: Route coordinates needed careful conversion and validation for accurate globe rendering.
Visual clarity at scale: Thousands of points and overlapping lines required iterative tuning of color, opacity, altitude, and render ordering.
Operational reliability: We handled process conflicts, stale server states, and local environment issues while keeping the demo stable.
Explainable recommendations: We wanted route recommendations to feel trustworthy, so we emphasized clear risk deltas, source context, and explicit fallback states.

Accomplishments that we're proud of

Built an end-to-end system from ingestion to action-oriented visualization
Added explainable AI-style route risk adjustments
Implemented safer-route highlighting that keeps endpoints consistent
Designed a polished interface that balances density with usability
Created a demo-ready workflow that communicates technical depth clearly

What we learned

Explainability is essential for operational AI tools.
UX details can make or break trust in technical systems.
Reliability and observability matter as much as model logic.
Building under hackathon constraints requires tight iteration loops and ruthless prioritization.

What's next for SupplyWatch

Integrate richer real-time maritime feeds (AIS, congestion, weather, port ops)
Improve optimization with multi-objective routing (risk, time, cost, reliability)
Add collaboration workflows (assignment, acknowledgement, escalation)
Expand historical analytics and forecasting for proactive planning
Productionize deployment, monitoring, and alerting for real enterprise use SupplyWatch is our step toward a true mission control layer for global supply chains.