Skip to content

niyatibalu/PULSE

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

73 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PULSE Logo

PULSE

Predictive Urban Logistics & Safety Engine

Turn historical crash data, live weather, and road conditions into actionable safety intelligence — with a built-in simulator to plan before harm occurs.


🌟 Overview

PULSE is a decision-support system for urban safety and EMS resource planning. It doesn't just tell you where accidents are likely — it helps city planners, dispatchers, and analysts figure out what to do about it, and what happens if they do something different.


💡 Problem

Most safety tools stop at the heatmap. PULSE is designed around the next question every operator actually asks:

"So what do we do about it?"


⚙️ How It Works

  1. 📊 Risk Forecasting — Machine learning models trained on Madison crash data, weather, and road conditions generate short-term risk predictions across Madison, gridded by location and time.

  2. 🚑 EMS Deployment Recommendations — Based on current forecasts, PULSE suggests optimal pre-positioning of limited emergency resources to minimize expected response time and harm.

  3. 💬 Explainable Outputs — Every forecast comes with a plain-language summary of the driving factors: weather, time of day, traffic volume, recent incidents.


🧑‍💻 Tech Stack

Data

  • WisDOT Crash Records (statewide, severity + coordinates)
  • NOAA / Open-Meteo (hourly weather history)
  • OpenStreetMap via OSMnx (road network)
  • US Census ACS (demographics, income, vehicle access)
  • Madison Open Data (fire stations, EMS locations)

Program

  • scikit-learn / XGBoost — crash risk model
  • OR-Tools — facility location & resource allocation
  • PyDeck — data visualization
  • Python, FastAPI
  • PostGIS for spatial queries
  • React + Leaflet (interactive risk map)
  • Recharts (confidence intervals, scenario comparisons)

🚀 Roadmap to Scaling

  • Real-time 911 dispatch integration
  • Multi-city deployment beyond Madison
  • Reinforcement learning for adaptive EMS routing
  • Public-facing risk transparency dashboard

Built for MadData 2026 — Madison, WI
By Abhinav, Niyati, Sahana

About

MadData Hackathon '26

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages