Student Depression Risk Prediction System

A comprehensive machine learning application that predicts depression risk in students using XGBoost algorithm with 85.2% accuracy. The system provides both Turkish and English interfaces for analyzing student mental health data.

🚀 Features

• Bilingual Support: Available in both Turkish and English
• Interactive Dashboard: Multi-page application with comprehensive analysis
• Data Exploration: Visualize dataset distributions and statistical insights
• Model Performance: Display accuracy metrics and feature importance analysis
• Risk Prediction: Interactive form to predict depression risk for individual students
• Educational Content: Step-by-step explanation of the ML development process
• Professional Recommendations: Evidence-based suggestions for intervention

📁 Project Structure

ML-Student-Depression/
├── streamlit_app.py              # English Streamlit application
├── streamlit_tr.py               # Turkish Streamlit application (Türkçe)
├── requirements.txt              # Python dependencies
├── main.ipynb                    # Original model development notebook
├── no_suicidal.ipynb            # Alternative analysis notebook
├── Student Depression Dataset.csv # Training dataset
├── xgboost_depression_model.pkl  # Trained XGBoost model
├── xgboost_scaler.pkl           # Feature scaler for model
├── scaler.pkl                   # Additional scaler file
├── feature_names.pkl            # Model feature names
└── README.md                    # Project documentation

🔧 Installation & Setup

1. Clone the repository:

   git clone <repository-url>
   cd ML-Student-Depression

2. Install dependencies:

   pip install -r requirements.txt

3. Run the Streamlit app:

   For English version:

   streamlit run streamlit_app.py

   For Turkish version (Türkçe):

   streamlit run streamlit_tr.py

4. Open your browser and navigate to:

   http://localhost:8501
   

📊 Application Sections

Both applications (English and Turkish) contain the same comprehensive sections:

🏠 About Project / Proje Hakkında

• Project overview and objectives
• Model accuracy and dataset statistics
• Key features and system capabilities

� How I Did It? / Nasıl Yaptım?

• Step-by-step development methodology
• Technical explanations made accessible
• Technology stack and tools used
• Model selection and optimization process

📊 Data Analysis / Veri Analizi

• Dataset overview and statistics
• Interactive data visualizations
• Depression distribution analysis
• Key statistical findings

🤖 Model Performance / Model Performansı

• Detailed performance metrics (Accuracy, Precision, Recall, F1-Score)
• Feature importance visualization
• Model comparison analysis
• Algorithm explanations and working principles

🔮 Risk Prediction / Risk Tahmini

• Interactive form for student data input
• Real-time depression risk assessment
• Risk level classification with recommendations
• Visual probability gauge and detailed explanations

📈 Findings and Recommendations / Bulgular ve Öneriler

• Key research findings and insights
• Evidence-based recommendations
• Institutional and individual guidance
• Future development possibilities

🎯 Model Information

• Algorithm: XGBoost Classifier (Extreme Gradient Boosting)
• Training Accuracy: 85.2%
• Dataset Size: 27,000+ student records
• Features: 16 engineered features including psychological, academic, and lifestyle indicators
• Risk Categories: Low 🟢, Medium 🟡, High 🟠, Very High 🔴
• Model Files: Pre-trained model, scaler, and feature mappings included

🔍 Key Predictive Features

The model analyzes multiple factors to assess depression risk:

1. Psychological Indicators

   • Suicidal thoughts history (highest impact factor)
   • Family mental health history
   • Overall life satisfaction

2. Stress Factors

   • Financial stress levels
   • Academic pressure
   • Work-life balance

3. Lifestyle Variables

   • Sleep duration patterns
   • Dietary habits
   • Daily study/work hours

4. Academic Factors

   • Study satisfaction
   • Grade point average (CGPA)
   • Degree level

5. Demographic Information

   • Age group (18-25 identified as high-risk)
   • Gender
   • Education level

🔬 Technical Details

• Development Environment: Python 3.8+ with Jupyter Notebooks
• ML Libraries: Scikit-learn, XGBoost, Pandas, NumPy
• Visualization: Plotly, Matplotlib, Seaborn
• Web Framework: Streamlit for interactive applications
• Model Persistence: Joblib for model serialization

🎨 Customization

You can customize the applications by:

• Styling: Modify CSS styles in the st.markdown() sections for custom themes
• Visualizations: Add new charts or modify existing plot configurations
• Risk Assessment: Update risk thresholds and recommendation logic
• Input Fields: Add new prediction parameters or modify existing form fields
• Language: Extend to other languages following the bilingual pattern
• Model Updates: Replace model files to use updated or different algorithms

🌐 Bilingual Support

This project demonstrates international accessibility:

• streamlit_app.py: Complete English interface
• streamlit_tr.py: Full Turkish interface (Türkçe arayüz)
• Both versions maintain identical functionality
• Easy to extend to additional languages
• Cultural sensitivity in mental health terminology

🚨 Important Notes

• File Dependencies: Ensure all model files (.pkl) are in the same directory
• Dataset: The CSV file must be named exactly "Student Depression Dataset.csv"
• Python Version: Requires Python 3.8 or higher
• Memory: XGBoost model requires sufficient RAM for large predictions
• Production: Add comprehensive data validation and error handling for deployment

🤝 Usage Guidelines

For Researchers and Developers:

• Use main.ipynb to understand the model development process
• Examine feature engineering techniques in the notebooks
• Adapt the methodology for similar mental health prediction tasks

For Educators and Institutions:

• Deploy the application for educational demonstrations
• Use risk predictions as supplementary screening tools
• Implement recommendations for institutional mental health programs

For Students and General Users:

• Navigate using the sidebar menu for different analysis sections
• Complete all fields in Risk Prediction for accurate results
• Remember that results are supplementary to professional assessment

� Ethical Considerations & Disclaimer

Important Medical Disclaimer:

• This tool is designed for educational and research purposes only
• Results should never replace professional mental health assessment
• Always consult qualified mental health professionals for diagnosis and treatment
• The system provides risk indicators, not definitive medical diagnoses

Ethical AI Use:

• Model predictions should supplement, not replace, human judgment
• Consider potential biases in training data and model decisions
• Maintain privacy and confidentiality of student information
• Use predictions responsibly within appropriate institutional frameworks

Data Privacy:

• No user input data is stored or transmitted
• All processing occurs locally within the application
• Follow institutional guidelines for handling student mental health data

📞 Troubleshooting & Support

Common Issues:

1. Missing Dependencies:

   pip install --upgrade streamlit plotly scikit-learn xgboost pandas numpy joblib seaborn matplotlib

2. Model File Errors:

   • Verify all .pkl files are in the same directory as the Python scripts
   • Check file permissions and accessibility

3. Dataset Issues:

   • Ensure CSV file is named exactly "Student Depression Dataset.csv"
   • Verify file encoding (UTF-8 recommended)

4. Performance Issues:

   • Close other applications to free memory
   • Consider using a virtual environment
   • Update to the latest Python version (3.8+)

Getting Help:

• Check GitHub Issues for known problems
• Review Streamlit documentation for deployment issues
• Verify system requirements and compatibility

🚀 Future Enhancements

Planned Features:

• Real-time monitoring dashboard
• Integration with learning management systems
• Mobile-responsive design
• API endpoints for institutional integration
• Advanced visualization options
• Multi-language support expansion

Research Opportunities:

• Longitudinal tracking capabilities
• Integration with wearable device data
• Social media sentiment analysis
• Intervention effectiveness tracking
• Cross-cultural validation studies

---

Project Status: ✅ Complete and Ready for Use
Last Updated: June 2025
License: Educational Use
Developed for: ML Student Depression Analysis Project