Student Performance Prediction

---

Overview

This project focuses on predicting academic performance among secondary school students using machine learning. The goal is to:

• Predict students' Grade Point Average (GPA)
• Classify students as at-risk based on predicted GPA (below a threshold)
• Compare classification model performance
• Uncover key factors that influence student success

This project empowers educators with early warnings and data-driven insights by combining academic, behavioral, and demographic features.

---

Dataset

The dataset contains a range of features such as:

• Demographics (e.g., gender, ethnicity, age)
• Academic indicators (e.g., study time, absences, parental education)
• Behavioral traits (e.g., extracurriculars, volunteering, tutoring)

📄 A full breakdown of features, data types, and example values can be found in the data catalog under the docs/ folder.

---

Features / Objectives

• GPA Prediction: Estimate students' GPA using regression techniques
• At-Risk Classification:
  • GPA-based risk flagging (Flag students as "At Risk" when the predicted GPA is below 2.0)
  • Direct classification using trained classifiers
  • Model comparison view included in the Streamlit app
• Feature Importance: Analyze key factors driving academic performance
• Streamlit App: Provide a simple user interface for live predictions
• Insights for Educators: Offer actionable data to improve student outcomes

---

Installation & Setup

1. Clone the Repository

git `clone` https://github.com/mandele1999/student_performance_project.git
cd student_performance_project

2. Install Required Packages

pip install -r requirements.txt

3. Launch the Streamlit App

cd scripts
streamlit run streamlit_app.py

📁 Project Structure

student_performance_project/
│
├── data/
│   ├── raw/                          
│   ├── processed/                   
│   └── feature_engineered_student_data.csv
│
├── docs/                                                 
│   ├── data_catalog.md
│   ├── app_guide.md
|   ├── model_report.md
│   └── project_overview.md
│
├── models/
│   ├── logreg_risk_classifier_model.pkl
│   ├── lr_student_grade_model.pkl
│   └── preprocessor_pipeline.pkl
│
├── notebooks/
│   ├── 01_data_preprocessing.ipynb  
│   ├── 02_feature_engineering.ipynb 
│   └── 03_modeling.ipynb            
│
├── scripts/
│   ├── custom_transformers.py
│   ├── inference.py
│   └── streamlit_app.py         
│   
├── requirements.txt
├── README.md
├── .gitignore
└── LICENSE

Usage (Streamlit App)

Once the app is running:

• Enter student details (age, study time, absences, etc.)
• Predict GPA using a trained Linear Regression model
• See two risk classification results:
  • Based on predicted GPA threshold
  • From a trained classifier model
• Compare model outputs in a single view
• Get instant feedback on potential student performance

Results/Insights

• Best Model: Linear Regression for GPA prediction
• Risk Classification: Compare both GPA-threshold-based and trained classifier models
• Top Influencers: Parental education, study time, absences, and support systems emerged as significant predictors
• App Feedback: Real-time, easy-to-understand results to assist decision-makers in schools

🤝 Contributing

Contributions are welcome! If you'd like to add improvements, raise issues, or contribute new ideas (e.g., dashboard features, model tuning, or alternative algorithms):

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

License

This project is licensed under the MIT License.

Acknowledgements

• Open-source Python tools and libraries
• Contributors and mentors from the data science community
• Student Performance Dataset on Kaggle