Sustainability Pro

Predicting pollution. Protecting people. Building a cleaner NYC.

Comment

Inspiration When New York City briefly held the title of worst air quality in the world during the 2023 Canadian wildfires, we realized how invisible — yet urgent — the issue of urban pollution truly is. We wanted to take action using what we know best: data and technology. Inspired by this crisis, we set out to transform air quality data into meaningful insights that empower communities to breathe cleaner, live healthier, and make smarter environmental decisions.

What it does Sustainability Pro is an AI-powered platform that visualizes, predicts, and educates users about urban air quality.

  • It combines real-time AQI data, machine-learning-based forecasts, and geographic analysis to:
  • Map pollution trends across NYC boroughs
  • Predict future pollutant levels using XGBoost models
  • Identify high-risk areas through composite pollution scoring
  • Recommend targeted actions — from green transit routes to policy strategies
  • Educate users through an interactive website that turns complex data into accessible insights In short, Sustainability Pro bridges the gap between environmental data and public awareness — turning numbers into action.

How we built it We began by collecting and cleaning over 13,000 data points from NYC’s Environment & Health Data Portal, including variables such as borough IDs, pollutant indicators, and hospitalization rates. Using Python and XGBoost, we trained gradient-boosted models to predict future pollutant concentrations (PM2.5, NO₂, and O₃). We then built an interactive website using Flask (or optionally React/Supabase) that integrates these predictions with live AQI data APIs. The frontend displays:

  • Pollution heat maps
  • Borough-level air quality predictions
  • Dynamic visualizations showing how trends evolve over time We also used a simple REST API to fetch data and feed predictions directly into the visualization dashboard.

Challenges we ran into

  • Cleaning and preprocessing inconsistent city-level environmental datasets
  • Tuning XGBoost hyperparameters to minimize prediction error without overfitting
  • Integrating multiple APIs (for live AQI data, maps, and historical archives) into one cohesive platform
  • Designing a frontend that conveys complex environmental metrics in a user-friendly way

Accomplishments that we're proud of

  • Successfully predicted NYC pollutant trends through machine-learning models with strong accuracy
  • Built an intuitive platform that connects technical insights to real-world community impact
  • Identified actionable policy zones for targeted congestion pricing and filtration solutions
  • Combined technical rigor, public health data, and design into one cohesive sustainability project

What we learned We learned that AI can play a crucial role in environmental sustainability — but only when combined with accessibility and community engagement. We deepened our understanding of:

  • How gradient boosting models work and how to interpret feature importance
  • The importance of open-source government data
  • Translating technical results into meaningful social change

What's next for Sustainability Pro We plan to:

  • Expand the platform to cover all major U.S. cities
  • Add a mobile app that alerts users of high-pollution areas in real time
  • Integrate IoT sensor data from public and private sources for improved accuracy
  • Partner with environmental agencies to pilot our policy simulator, showing how interventions like congestion pricing or electric transit could change future air quality Our vision is simple: Empower every city — and every citizen — to make data-driven decisions for a cleaner future.

Built With

Share this project:

Updates