🏎️ Formula 1 Qualifying Lap Time Predictor (FastF1 + XGBoost)

Predict F1 Qualifying Lap Times and Grid Order using Machine Learning

This project uses FastF1 telemetry data and a gradient boosting model (XGBoost) to predict qualifying lap times for Formula 1 races — including a simulated 2025 United States GP (Austin) qualifying session.

---

📚 Overview

This repository demonstrates how to:

• Collect historical qualifying data (2021–2023) from FastF1
• Merge weather and track conditions (air temp, track temp, humidity, etc.)
• Train a machine learning model to predict lap times
• Generate a predicted qualifying order for any event
• Visualize feature importance, model performance, and qualifying results

---

🧠 Project Highlights

✅ Uses real F1 telemetry via the FastF1 API
✅ Predicts realistic lap times (~1:34 for Austin)
✅ Model achieves R² ≈ 0.99 and MAE ≈ 0.3 s
✅ Includes weather-aware predictions
✅ Generates F1-style timing sheets with gaps (+0.xxx)
✅ Extensible for “what-if” scenarios (e.g., driver/team swaps or temperature changes)

---

🧩 Example Output

🏁 Predicted 2025 Austin GP Qualifying Order:

## Pos  Driver  Team                    Lap Time     Gap

1    SAI    Ferrari                1:34.147     +0.000
2    VER    Red Bull Racing        1:34.285     +0.138
3    LEC    Ferrari                1:34.473     +0.325
4    PER    Red Bull Racing        1:34.557     +0.410
5    HAM    Mercedes               1:34.743     +0.596
...

📈 Model Performance:
MAE = 0.321 s | R² = 0.988
🌡️ Austin Conditions: Air 32°C | Track ≈ 42°C

---

📊 Model Details

Features Used:

• Driver
• Team
• Event (Circuit)
• Tyre Compound
• Tyre Life
• Air Temperature
• Track Temperature
• Humidity
• Pressure
• Wind Speed
• Track Status

Target:

• Lap time (normalized per event)

Algorithm:

• XGBRegressor

  • 800 estimators
  • learning_rate = 0.05
  • max_depth = 7
  • subsample = 0.8
  • colsample_bytree = 0.8

---

🔬 Insights You Can Explore

• Driver vs. teammate performance over multiple seasons
• Circuit clusters (fast vs. technical tracks)
• Weather sensitivity — effect of air/track temp on lap time
• Team development trends year over year
• What-if simulations (e.g., “What if Leclerc drove a Red Bull?”)

---

🏁 Credits

• Data: FastF1 (OpenTelemetry API)
• Weather: Open-Meteo API
• Modeling: XGBoost & Scikit-learn
• Visualization: Matplotlib

---

📜 License

This project is released under the MIT License. Feel free to fork, modify, and build your own F1 analytics models!

---

“Data wins races — if you know how to read it.” 🏁