🚕 NYC Taxi Trip Duration Prediction — Production ML Service

---

📌 Problem

• Taxi trip duration in NYC varies significantly by time, location, and traffic
• Inaccurate estimates impact passenger ETAs, fleet utilization, and pricing decisions
• Rule-based or manual estimation does not scale to city-level operations

---

🧠 Approach

• Trained a regression model on historical NYC Taxi trip data
• Built an API-first inference service using FastAPI
• Enforced schema validation to block invalid inputs
• Reused the same preprocessing pipeline at training and inference (no train–serve skew)
• Deployed as a Dockerized service for reproducibility and consistency

---

📊 Metric

• Target: Trip duration (minutes)
• Evaluated using regression metrics (MAE, RMSE) on held-out data
• Focused on stable, reliable predictions rather than overfitting for benchmark scores

---

💼 Business Impact

• Enables real-time ETA estimation for passengers
• Supports driver allocation and fleet planning
• Helps detect abnormal or inefficient trips
• Architecture reusable for other prediction services (ETA, demand, churn, fraud)

---

🚀 Key Takeaway

Transforms real-world NYC taxi data into a reliable, production-ready machine-learning prediction service.

---

Workflow

⚡ Quickstart

# Create virtual environment 
python -m venv venv
source venv/bin/activate   # Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Start API
uvicorn app.main:app --reload
# Open
http://localhost:8000