🏎️ Formula 1 Race Winner Prediction Model

A comprehensive machine learning project that predicts Formula 1 race winners using historical data from 1950 to 2024. The model leverages 52+ engineered features from multiple data sources including race results, qualifying times, pit stops, lap times, sprint races, and championship standings.

📋 Table of Contents

• Overview
• Dataset

🎯 Overview

This project builds and compares 10 different machine learning models to predict Formula 1 race winners. The system automatically selects the best-performing model based on comprehensive evaluation metrics including accuracy, F1-score, precision, recall, and ROC-AUC.

Key Highlights

• Comparision of 10 Machine Learnining Models: Random Forest, XGBoost, Gradient Boosting, Extra Trees, AdaBoost, Decision Tree, Logistic Regression, KNN, Naive Bayes, and SVM
• 52+ Engineered Features: Comprehensive feature engineering from 12 different data sources
• Time-Based Validation: Proper temporal split to prevent data leakage
• Automatic Best Model Selection: Identifies the optimal model based on test performance
• Real-World Application: Can predict winners for future races given driver and race data

📊 Dataset

The dataset is sourced from Kaggle - Formula 1 World Championship (1950-2024), compiled from Ergast API.

Data Sources Used

The model integrates 12 CSV files:

1. circuits.csv - Circuit information (location, country, coordinates)
2. constructor_results.csv - Constructor race results
3. constructor_standings.csv - Constructor championship standings
4. constructors.csv - Constructor/team details
5. driver_standings.csv - Driver championship standings
6. drivers.csv - Driver information (nationality, DOB)
7. lap_times.csv - Lap-by-lap timing data
8. pit_stops.csv - Pit stop strategy and duration
9. qualifying.csv - Qualifying session times (Q1, Q2, Q3)
10. races.csv - Race schedule and metadata
11. results.csv - Final race results and positions
12. sprint_results.csv - Sprint race results
13. status.csv - Race finish status (completed, DNF, etc.)

🔧 Features

Feature Categories (52+ total features)

1. Grid & Qualifying Features

• Starting grid position
• Qualifying position (final)
• Q1, Q2, Q3 lap times
• Best qualifying time

2. Sprint Race Features

• Sprint race position
• Sprint points earned
• Sprint grid position
• Sprint race participation flag

3. Lap Time & Performance Features

• Average lap time
• Lap time standard deviation (consistency)
• Fastest/slowest lap times
• Total laps completed
• Lap time consistency metric

4. Pit Stop Strategy Features

• Number of pit stops
• Average pit stop duration
• Total time lost in pits
• Fastest/slowest pit stop

5. Driver Historical Performance

• Previous wins, podiums, top 5, top 10 finishes
• Total career points
• Average finishing position
• Recent form (last 5 and 10 races)
• Win rate and podium rate
• DNF (Did Not Finish) rate
• Average starting grid position
• Average qualifying position

6. Constructor/Team Performance

• Team's previous wins and podiums
• Team's total points
• Team's average finishing position
• Team's recent form
• Team win rate and podium rate

7. Circuit-Specific Performance

• Driver's races at specific circuit
• Driver's wins at specific circuit
• Driver's podiums at specific circuit
• Driver's average position at circuit
• Circuit-specific win rate

8. Championship Context

• Championship points before race
• Championship position before race
• Championship wins before race

9. Encoded Categorical Features

• Driver ID (encoded)
• Constructor ID (encoded)
• Circuit ID (encoded)