FIFA Player Rating Prediction

A comprehensive machine learning project that predicts FIFA player overall ratings using advanced ensemble methods and provides an interactive web interface for real-time predictions.

🚀 Features

• Advanced ML Pipeline: Complete data preprocessing with imputation and standardization
• Ensemble Methods: Random Forest, XGBoost, and Gradient Boosting regressors
• Hyperparameter Tuning: Automated optimization using RandomizedSearchCV
• Cross-Validation: Robust model evaluation with 3-fold cross-validation
• Interactive Web App: Streamlit-based interface with visual star ratings
• Model Persistence: Trained models saved using joblib for deployment

📊 Dataset

The project uses FIFA player datasets:

• Training Data: male_players (legacy).csv
• Testing Data: players_22-1.csv (different season for validation)

Key Features Used

• Movement Reactions
• Mentality Composure
• Passing & Dribbling
• Physical Attributes
• Shooting & Shot Power
• Age and other performance metrics

🛠️ Technology Stack

• Python 3.7+
• Machine Learning: scikit-learn, XGBoost
• Data Processing: pandas, numpy
• Model Persistence: joblib
• Web Interface: Streamlit
• Visualization: Built-in Streamlit components

📋 Installation

Prerequisites

pip install streamlit pandas scikit-learn joblib xgboost numpy scipy

Clone Repository

git clone https://github.com/Ama-Annor/AMAANNOR._SportsPrediction.git
cd AMAANNOR._SportsPrediction

Install Dependencies

pip install -r requirements.txt

🎯 Usage

1. Train the Model

Run the Jupyter notebook or Python script to train the model:

python AMAANNOR._SportsPrediction.py

This will:

• Clean and preprocess the data
• Train multiple ML models
• Perform hyperparameter tuning
• Save the best model as model_best.pkl
• Save the scaler as scaler.pkl

2. Launch the Web App

streamlit run player_rating_app.py

3. Make Predictions

1. Open your browser to the Streamlit app (usually http://localhost:8501)
2. Adjust player attribute sliders
3. Enter actual rating for confidence calculation
4. Click "Predict" to see results with star ratings

🧠 Machine Learning Pipeline

Data Preprocessing

1. Feature Selection: Focus on 11 most correlated features with player rating
2. Missing Value Imputation: Median-based imputation using SimpleImputer
3. Standardization: StandardScaler for feature normalization
4. Data Cleaning: Remove non-numeric columns and handle missing values

Model Training & Evaluation

# Models used
- Random Forest Regressor
- XGBoost Regressor  
- Gradient Boosting Regressor (Best performing)

# Evaluation Metrics
- Mean Absolute Error (MAE)
- Root Mean Squared Error (RMSE)
- R² Score
- Cross-validation scores

Hyperparameter Optimization

• Method: RandomizedSearchCV
• Parameters: n_estimators, learning_rate, max_depth, min_samples_split, etc.
• Cross-validation: 3-fold CV
• Scoring: Negative Mean Squared Error

📈 Model Performance

Cross-Validation Results

• Random Forest: Mean CV Score: ~0.85
• XGBoost: Mean CV Score: ~0.87
• Gradient Boosting: Mean CV Score: ~0.89 (Best)

Final Model Metrics

• RMSE: ~0.15-0.20 (on scaled data)
• MAE: ~0.12-0.18
• R² Score: ~0.89-0.92

🌟 Web Application Features

Interactive Interface

• Slider Controls: Easy adjustment of player attributes
• Real-time Predictions: Instant rating calculation
• Visual Feedback: Star-based rating system (⭐⭐⭐⭐⭐)
• Confidence Metric: Shows prediction reliability

Star Rating System

def compute_stars(rating):
    # Converts numerical rating to 5-star scale
    # Includes half-stars and visual representations

📁 Project Structure

AMAANNOR._SportsPrediction/
├── AMAANNOR._SportsPrediction.ipynb    # Main training notebook
├── AMAANNOR._SportsPrediction.py       # Python script version
├── player_rating_app.py                # Streamlit web application
├── model_best.pkl                      # Trained model (generated)
├── scaler.pkl                          # Data scaler (generated)
├── requirements.txt                    # Python dependencies
├── male_players (legacy).csv           # Training dataset
├── players_22-1.csv                    # Testing dataset
└── README.md                           # Project documentation

🔧 Configuration

Model Parameters (Optimized)

GradientBoostingRegressor(
    n_estimators=300,
    learning_rate=0.1,
    max_depth=6,
    min_samples_split=5,
    min_samples_leaf=2,
    subsample=0.8
)

📊 Results & Insights

Feature Importance

Top contributing factors to player ratings:

1. Movement Reactions
2. Mentality Composure
3. Passing Ability
4. Dribbling Skills
5. Physical Attributes

Model Validation

• Tested on different season data (players_22)
• Maintains consistent performance across datasets
• Robust to new, unseen player data

🚀 Future Enhancements

• Deep learning models (Neural Networks)
• Player position-specific models
• Time series analysis for rating changes
• Advanced feature engineering
• Model interpretability with SHAP values
• API deployment for mobile apps
• Real-time data integration

📝 Requirements

Create a requirements.txt file:

streamlit>=1.28.0
pandas>=1.5.0
scikit-learn>=1.3.0
joblib>=1.3.0
xgboost>=1.7.0
numpy>=1.24.0
scipy>=1.10.0

🤝 Contributing

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

👨‍💻 Author

Ama-Annor

• GitHub: @Ama-Annor

🙏 Acknowledgments

• FIFA for providing comprehensive player statistics
• scikit-learn community for excellent ML tools
• Streamlit team for the intuitive web framework
• XGBoost developers for high-performance gradient boosting

---

⚽ Predict like a pro with data science! 🏆