PropertyIQ 🏠🧠

A machine learning-powered house price prediction system for Winnipeg real estate, featuring data processing pipelines, ML model training, and an interactive web application.

🎯 Overview

The City of Winnipeg releases comprehensive sales data for houses and condominiums every 2 years in PDF format (Sales Books). This project transforms that raw data into actionable insights by:

• Extracting and processing PDF sales data into structured datasets
• Training ML models using scikit-learn to predict house prices based on property features
• Providing an intuitive web interface for real-time price predictions

The model leverages property characteristics from the Winnipeg Tax Assessment dataset including living area, land area, number of rooms, basement/pool presence, and more.

🏗️ Architecture

├── model/           # ML pipeline & data processing
├── webapp/          # React frontend application  
├── server/          # Flask API backend
└── README.md        # This file

Key Features

• PDF-to-CSV data extraction pipeline (using Tabula for PDF processing)
• Feature engineering and data preprocessing
• Trained scikit-learn regression model
• React TypeScript frontend with form validation
• Flask REST API for predictions
• Real-time price predictions

🚀 Quick Start

Prerequisites

• Python 3.8+ - Download here
• Node.js 16+ with pnpm - Install pnpm
• Git for version control

1. Clone Repository

git clone <repository-url>
cd repository

2. Train the Model

cd model/
# Follow instructions in model/README.md to process data and train the model

3. Start the Backend Server

cd server/
py -m pip install -r requirements.txt
py server.py

✅ Server running at http://localhost:5000

4. Start the Frontend

cd webapp/
pnpm install
pnpm dev --host

✅ Web app running at http://localhost:3000

5. Test the Application

Visit http://localhost:3000 in your browser and input house features to get price predictions!

📊 Data Sources

• Sales Data: City of Winnipeg biennial house/condo sales (PDF format, 2016-2021)
  • Note: Tabula was used externally to extract data from PDF sales books into CSV format for model consumption
• Property Features: Winnipeg Tax Assessment parcels dataset
• Coverage: Houses only (condos excluded from current model)

🛠️ Development Setup

Recommended VS Code Extensions

{
  "recommendations": [
    "ms-python.python",
    "ms-toolsai.jupyter", 
    "dbaeumer.vscode-eslint",
    "esbenp.prettier-vscode",
    "GraphQL.vscode-graphql"
  ]
}

Project Structure Details

• model/notebooks/ - Jupyter notebooks for data analysis and training
• model/datasets/ - Raw PDF data and processed CSV files
• webapp/src/ - React components and styling
• server/ - Flask API with prediction endpoint

📈 Model Performance

The model uses property features including:

• Living area (sq ft)
• Land area (sq ft)
• Number of rooms/bedrooms/bathrooms
• Basement/pool presence
• Property age and condition

For detailed performance metrics, see model/notebooks/train_model.ipynb

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit changes (git commit -m 'Add amazing feature')
4. Push to branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📝 License

This project was developed as part of a work term at Norima Technologies (January–April 2022). Please ensure compliance with City of Winnipeg data usage guidelines when using or distributing this repository.

---

Note: This model is for educational/demonstration purposes. Actual real estate valuations should involve professional appraisals and additional market factors.