Night Guard

NightGuard uses real FEMA, USGS, and climate data to generate your personal risk score, showing exactly what's driving it and how to lower it before a claim hits.

Comment

Inspiration Most people don’t understand their insurance risk until it’s too late. While FEMA and USGS data exist, they aren't accessible to average homeowners. We built NightGuard to surface this data and empower families to make informed coverage decisions.

What it does NightGuard is an AI risk assessment platform that blends government data (FEMA, USGS, Open-Meteo) with a personalized quiz to generate a Legacy Score. Unlike a "black-box" quote, we provide: Transparent Drivers: Labeled factors showing exactly what raised or lowered your score. Actionable Advice: Ranked recommendations to mitigate specific risks. Gamified Progress: Streaks, badges, and trajectory charts to track safety improvements. Live Mapping: Mapbox integration with FEMA flood zone overlays.

How we built it Frontend: React, Vite, Tailwind CSS, and Framer Motion for a polished, "museum-archive" aesthetic. Backend: FastAPI (Python) integrating FEMA, USGS, and Open-Meteo APIs. AI Layer: Groq (Llama 3.3 70B) for score generation and risk analysis, featuring a logic-based fallback system to ensure scores are never hardcoded. Auth: Supabase for secure user authentication.

Challenges & Wins The Pivot: We hit API limits on other models mid-hackathon and successfully migrated our entire scoring pipeline to Groq. Transparency: We moved away from "black-box" AI by designing a weighting system that visualizes exactly how user inputs impact their score. Data Integrity: We’re proud of successfully piping four live government data sources into a single, cohesive dashboard.

What's next Mobile App: developing a mobile version of this for on-the-go claims checking Real-time Alerts: Push notifications for local hazards via OpenWeatherMap. Integrations: Adding Zillow API for property context and a mobile app for on-the-go alerts.

Built With

Share this project:

Updates