🏈 NFL Super Bowl Prediction (2003–2023)
This project predicts the probability of an NFL team winning the Super Bowl based on their regular season performance using machine learning.
We dive deep into over 20 years of NFL team stats, engineer meaningful features, train multiple models, and test our predictions. 



🔍 What This Project Does:
- Analyzes NFL team season data from 2003 to 2023

- Engineers features like turnover percentage, score efficiency, and point differentials

- Trains a Random Forest Classifier to predict if a team will win the Super Bowl

- Calculates and ranks each team’s win probability

- Tests the model on new unseen season data (e.g. 2022)

- Visualizes trends and top predictions


✅ Which Model was Chosen?
- We chose Random Forest due to better performance on the minority class (Super Bowl winners).


📈 Visualizations
- Distribution of win percentages

- Heatmap of feature correlations

- Top 10 predicted teams for Super Bowl (by win probability)


🔬 Future Improvements
- Use better dataset with less headings

- The data in the dataset is too much to get in season for proper modelling of the the year's winners using that year's regular season stats.

- Integrate playoff performance

- Find new ways like web apps to display the findings (eg a web based dashboard)

🧠 Built With
- Python, Pandas, Scikit-learn, Seaborn, Matplotlib

- Jupyter Notebooks