Aug 8, 2026

AIRAA

ai routing nlp geospatial women safety

AIRAA (Adaptive Intelligence for Risk Awareness & Action) is an AI-powered safety navigation system that helps women choose safer routes through a city, not just faster ones. It turns scattered, hard-to-trust safety information into a live map - and then, unlike any existing safety app, actually uses that map to route people around danger.

What it does:

  • Combines community-reported incidents, public safety data, and environmental factors (lighting, foot traffic) into one live risk map of a city
  • Uses AI to read and categorize free-text incident reports automatically - no manual tagging needed
  • Recommends walking routes that actively avoid high-risk areas, not just the fastest path
  • Filters out unreliable or spam reports before they can distort the map, using corroboration checks
  • Keeps individual reporters anonymous while still surfacing trustworthy, aggregate risk patterns
  • Shows why an area is flagged risky - not just a red zone with no explanation

Why it's different:

  • Most existing safety apps stop at showing pins on a map - AIRAA is the first to turn that data into an actual safer route
  • Built with trust and fairness as core features, not afterthoughts - every score is explainable and corroboration-checked before it counts
  • Designed to plug into real public safety data over time, not just user reports alone

Goal: Help someone get where they're going safely today, while giving city authorities a more honest, data-backed picture of where infrastructure like lighting and patrols are actually needed.

This build was uploaded as a hackathon project

Hackathon

Girls Hack Day in delhi

View All Projects

1

Give a star to encourage!Discussion
Start a new conversation!
Login to join the discussion
Updates
  • We identified and resolved a startup crash on our Render deployment that was causing the live demo to show a blank page. The backend now starts reliably, and the frontend correctly connects to the risk-scoring and routing APIs. What's working end-to-end now: • Incident report classification (Gemini + local ML fallback) • Live risk grid computation for the Chennai OMR pilot zone • Safety-weighted routing — shortest, safest, and alternative paths • Emergency SOS locator (nearby police stations & hospitals) Next up: adding an explainability panel so users can see exactly why an area is flagged as risky, and moving to persistent storage for report data.
    Saturday, Aug 8th, 2026