Airbnb Price Prediction - NYC

Machine learning model for predicting optimal nightly prices for Airbnb listings in New York City using property features and market data.

Overview

This project implements an end-to-end machine learning pipeline that predicts Airbnb listing prices with 97.2% accuracy (R² score). The system includes data processing, model training, evaluation, and deployment via both a web dashboard and REST API.

https://airbnb-pricing-etutc5pereohmu5bbtxdeo.streamlit.app/#evaluation-metrics

Key Features

• High Accuracy: 97.2% R² score on test set
• Production-Ready: Includes both dashboard and API interfaces
• Robust Pipeline: Automated data cleaning, feature engineering, and model training
• Interactive Dashboard: Streamlit-based UI for easy price predictions
• RESTful API: FastAPI service for programmatic access
• Comprehensive Evaluation: Detailed metrics and visualizations

Model Performance

Metric	Value
R² Score	0.972
RMSE	$49.19
MAE	$3.69
Median % Error	0.00%
Test Set Size	4,219 listings

Project Structure

airbnb-pricing/
├── data/                   # Data storage
│   ├── raw/               # Original datasets
│   └── processed/         # Cleaned data
├── models/                # Trained models
├── src/                   # Source code
│   ├── ingest.py         # Data loading
│   ├── clean.py          # Data cleaning
│   ├── features.py       # Feature engineering
│   ├── train.py          # Model training
│   ├── evaluate.py       # Model evaluation
│   └── predict.py        # Inference
├── app/                   # Streamlit dashboard
├── api/                   # FastAPI service
└── tests/                 # Unit tests

Quick Start

Installation

1. Clone the repository

git clone <repository-url>
cd airbnb-pricing

2. Create and activate virtual environment

python3 -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

4. Download data

• Visit Inside Airbnb
• Download NYC listings.csv.gz
• Place in data/raw/ directory

Training the Model

source .venv/bin/activate
cd src
python clean.py    # Clean data
python train.py    # Train model
python evaluate.py # Evaluate performance

Or use the automated pipeline:

./run_pipeline.sh

Using the Dashboard

./start_streamlit.sh

Open http://localhost:8501 in your browser.

Using the API

./start_api.sh

API documentation available at http://localhost:8000/docs

Usage Examples

Python API

from src.predict import predict_from_dict

listing = {
    'accommodates': 4,
    'bedrooms': 2,
    'bathrooms': 2.0,
    'room_type': 'Entire home/apt',
    'latitude': 40.7580,
    'longitude': -73.9855,
    'number_of_reviews': 50,
    'amenities': 'WiFi,Kitchen,Heating',
    'neighbourhood_cleansed': 'Manhattan',
    'property_type': 'Apartment',
    'bed_type': 'Real Bed',
    'cancellation_policy': 'moderate'
}

price = predict_from_dict(listing)
print(f"Predicted price: ${price:.2f}/night")

REST API

curl -X POST "http://localhost:8000/predict" \
  -H "Content-Type: application/json" \
  -d '{
    "accommodates": 2,
    "bedrooms": 1,
    "bathrooms": 1.0,
    "room_type": "Entire home/apt",
    "latitude": 40.7580,
    "longitude": -73.9855
  }'

Technical Details

Algorithm

• Model: Random Forest Regressor
• Preprocessing: StandardScaler for numeric features, OneHotEncoder for categorical
• Hyperparameter Tuning: RandomizedSearchCV with 5-fold cross-validation
• Training Data: 21,094 NYC listings

Features

Input Features:

• Property details (bedrooms, bathrooms, accommodates)
• Location (latitude, longitude, neighborhood)
• Policies (minimum nights, cancellation policy)
• Reviews and ratings

Engineered Features:

• Distance to city center
• Amenities count
• Temporal features (days since last review, host tenure)
• Interaction features (capacity × bedrooms, reviews × rating)

Dependencies

Core requirements:

• Python 3.8+
• pandas
• scikit-learn
• numpy
• matplotlib
• joblib

Web services:

• streamlit (dashboard)
• fastapi, uvicorn (API)

See requirements.txt for complete list.

Model Evaluation

The model includes comprehensive evaluation with:

• Performance metrics (RMSE, MAE, R²)
• Residual analysis
• Error distribution visualizations
• Comparison against baseline models

Run evaluation to generate detailed reports:

python src/evaluate.py

Results saved to models/residuals.png

Documentation

• MODEL_DOCUMENTATION.md: Complete model specifications and usage guide
• QUICKSTART.md: Quick reference for common commands
• PROJECT_SUMMARY.md: Detailed project architecture

Testing

pytest tests/

Production Deployment

Docker (Optional)

docker-compose up

Deployment Considerations

• Model size: 137 MB
• Inference time: <100ms
• Memory usage: ~500MB
• Recommended: Retrain quarterly for NYC market

Limitations

• Trained specifically for NYC market
• Best performance for standard listings ($50-$500/night)
• Static model (requires retraining to capture seasonal trends)
• May underperform on luxury or unique properties

Data Source

Data provided by Inside Airbnb, a mission-driven project providing data about Airbnb's impact on residential communities.

Contributing

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

License

This project uses publicly available data from Inside Airbnb. Review their data policies before commercial use.

Acknowledgments

• Inside Airbnb for providing open data
• NYC Open Data for geographic information
• Scikit-learn team for machine learning tools

Contact

For questions or issues, please open an issue on GitHub or refer to the documentation files.

---

Note: This is an educational/research project. Always verify predictions with current market data before making business decisions.