Reveal82

This repository contains the front-end for the Reveal82 project, a civic data science initiative to help Chicago residents understand and mitigate lead exposure risks in their water service lines.

🚀 Live Demo

Explore the Reveal82 web app

📋 Table of Contents

• Overview
• Team
• Repository Structure
• Features
• Tech Stack
• Setup & Installation
• Usage
• Future Plans
• Acknowledgements

Overview

Reveal82 is an interdisciplinary IPRO project combining statistical analysis, machine learning models, and an interactive front-end to predict and visualize lead service line risks across Chicago neighborhoods. Users can check their address risk score, explore spatial regression maps, and access recommendations for remediation.

Team

• Eileen Garay (Co-Leader / Statistical team)
• Lalith Kothuru (Neural Networks, Full-Stack dev)
• Elijah Perez (Data Merging, EDA, Random Forest)
• Virginia Reider (Co-Leader / Statistical Lead)
• Austin Samuel (Data Wrangler / ML)
• Rajan Savani (XGBoost, ML Lead)

Repository Structure

• Lead-documents: Contains all of the lead documents used in our analyses.
• ML: Contains the ML team’s data-merging scripts, notebooks, and trained models.
• Stats: Contains the Statistics team’s RMarkdown files, spatial-regression scripts, and visualizations.
• website: Contains the Next.js front-end source code for the Reveal82 website.
• README.md: This document.

├── Lead-documents/
│   Contains all of the lead documents used in our analyses.
├── ML/
│   Contains the ML team’s data-merging scripts, notebooks, and trained models.
├── Stats/
│   Contains the Statistics team’s RMarkdown files, spatial-regression scripts, and visualizations.
├── website/
│   Contains the Next.js front-end source code for the Reveal82 website.
└── README.md
    This document.

Features

• Check Your Risk: Enter any Chicago address to view lead risk scores, model details, and remediation resources.
• Spatial Maps: Interactive neighborhood‐level maps showing predicted lead concentration percentiles and service line material probabilities.
• Model Insights: Overview of model performance metrics, feature importances, and decision boundaries.
• Active Learning: List of high‐uncertainty addresses for prioritized testing and feedback loop.

Tech Stack

• Front-end: Next.js, Tailwind CSS
• APIs: Google Maps Places API for address lookup
• Models & Data: Jupyter notebooks (Rmd, .ipynb) for regression and ML workflows, Pandas, scikit-learn, XGBoost, TensorFlow/Keras
• Deployment: Vercel for front-end, GitHub for version control

Setup & Installation

1. Clone the repo

   git clone https://github.com/LALITH0110/reveal82website.git
   cd reveal82website

2. Install dependencies

npm install

3. Environment variables

• Create a .env.local with your Google Maps API key:

NEXT_PUBLIC_GOOGLE_MAPS_API_KEY=YOUR_API_KEY

4. Run locally

npm run dev

5. Visit http://localhost:3000

Usage

• Navigate to Check Your Risk, input an address, and submit.
• Explore the Data & Analysis tab for spatial map overlays and model insights.
• Use the Resources section for lead service line replacement guidance.

Future Plans

• Finalize neighborhood‐level regression visualizations and survival/time‐series analyses.
• Integrate real‐time testing feedback via active learning API.
• Enhance UI with filter/sort controls and additional model explanations.

Acknowledgements

• City of Chicago Open Data for service line and assessor datasets.
• BlueConduit for inspiration on active learning workflows.