UNREACHABLE: AI-Powered Emergency Accessibility Intelligence Platform
Submission for Hacktoberfest Open-Source AI Challenge — Week 2: Side Hustle What I Built UNREACHABLE addresses a critical gap in emergency response planning: knowing a building’s location does not guarantee that an emergency vehicle can reach it. The platform aims to identify road bottlenecks, asse
Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle Submission 💸
Submission for Hacktoberfest Open-Source AI Challenge — Week 2: Side Hustle
UNREACHABLE addresses a critical gap in emergency response planning: knowing a building’s location does not guarantee that an emergency vehicle can reach it.
The platform aims to identify road bottlenecks, assess vehicle-specific accessibility, analyze street images for potential obstructions, and estimate the impact of infrastructure improvements.
Potential users include emergency services, municipal authorities, and urban planners who need better information to prioritize access issues.
[Watch the demo video] http://drive.google.com/file/d/1WSaQ2wPVNFKtvxH-kamM_aK3rJC_LmAl/view
[Explore the GitHub repository] https://github.com/shiriei/UNREACHABLE
UNREACHABLE combines a web-based mapping interface with geospatial analysis and an AI-assisted visual inspection workflow. The planned architecture uses React and TypeScript for the frontend, FastAPI for backend services, GeoJSON for spatial data, and OpenAI vision capabilities to identify potential obstructions in street photographs.
The system is designed to distinguish verified information from uncertain findings, keeping human review central to safety-critical decisions.
Emergency accessibility is a shared infrastructure problem, not a problem any single organization can solve alone.
Open innovation enables developers, researchers, and civic authorities to improve mapping data, validate detection methods, and adapt the platform to local road conditions. It also makes the approach more transparent, extensible, and accessible to communities that need it most.