PotholeIQ NYC

Inspiration

The idea for PotholeIQ was inspired by the thought that, just as the Earth is always lit up somewhere, urban issues like potholes are ever-present and require constant attention. We noticed a growing trend in NYC around pothole discussions, especially with the mayor’s plan to fill 500,000 potholes, and wanted to create a smarter, data-driven solution.

What it does

PotholeIQ provides a real-time risk intelligence map for NYC potholes. It tracks new pothole reports, scores their risk and urgency using machine learning, and visualizes them on an interactive map. The platform helps prioritize repairs and alerts the NYC DOT to the most critical issues.

How we built it

Aggregated NYC Open Data (311, traffic, AADT, collisions) Engineered features and trained XGBoost models for risk and urgency Built a FastAPI backend with SQLite for serving predictions and map data Developed a modern frontend with React, Vite, and Tailwind CSS Deployed using Docker and Render.com

Challenges we ran into

Cleaning and joining disparate city datasets Handling missing or inconsistent data (e.g., NaN traffic counts) Ensuring model generalization without ground-truth labels Building a responsive, intuitive map UI under hackathon time constraints

Accomplishments that we're proud of

End-to-end pipeline from raw data to live map Real-time risk scoring and automated alerts Seamless integration of multiple public datasets A user-friendly, visually appealing frontend

What we learned

We learned how to integrate real-time public data, apply machine learning to urban problems, and deploy scalable cloud solutions. We also gained experience in rapid prototyping and cross-functional teamwork.

What's next for PotholeIQ NYC

We plan to expand to other cities, incorporate more data sources (like weather and road sensors), and partner with city agencies for automated repair dispatch. Our goal is to make urban infrastructure safer and smarter, one pothole at a time.