EduShield AI is an end-to-end Machine Learning and Explainable AI (XAI) web application designed to identify students at risk of academic dropout. By combining powerful ensemble learning models with SHAP-based explainability, the system provides accurate predictions along with transparent insights into the factors influencing each prediction.
🌐 Web Application
https://edushield-ai.streamlit.app/
Student dropout remains a major challenge for educational institutions. Early identification of at-risk students enables educators to provide timely interventions and improve academic outcomes.
EduShield AI predicts a student's academic status using demographic, socioeconomic, and academic information, while also explaining the reasoning behind each prediction using Explainable Artificial Intelligence (XAI).
- 🎓 Graduate
⚠️ Dropout- 📚 Enrolled
- 📊 Interactive Dashboard
- 📁 Dataset Explorer
- 📈 Exploratory Data Analysis (EDA)
- 🤖 Multiple Machine Learning Models
- 📉 Model Performance Comparison
- 🎯 Student Risk Prediction
- 🔍 SHAP Explainability
- 📊 Confusion Matrix & ROC Curve
- 📥 Download Prediction Results
- ☁️ Cloud Deployment with Streamlit
The application compares multiple supervised learning algorithms:
- Logistic Regression
- Decision Tree
- Random Forest
- XGBoost
- LightGBM
Each model is evaluated using standard classification metrics including Accuracy, Precision, Recall, F1-score, and ROC-AUC.
EduShield AI integrates SHAP (SHapley Additive exPlanations) to improve model transparency.
The dashboard provides:
- Global Feature Importance
- Local Prediction Explanation
- SHAP Summary Plot
- Individual Feature Contribution Analysis
This enables educators to understand why a particular prediction was made instead of relying on a "black-box" model.
Dataset Name
Predict Students' Dropout and Academic Success
Source
UCI Machine Learning Repository
https://archive.ics.uci.edu/dataset/697/predict+students+dropout+and+academic+success
- 4,424 student records
- 36 input features
- 3 target classes
Features include:
- Academic Performance
- Previous Qualifications
- Demographic Information
- Family Background
- Financial Information
- Enrollment Information
| Category | Technology |
|---|---|
| Programming Language | Python |
| Framework | Streamlit |
| Data Processing | Pandas, NumPy |
| Visualization | Matplotlib, Plotly |
| Machine Learning | Scikit-learn, XGBoost, LightGBM |
| Explainable AI | SHAP |
| Deployment | Streamlit Community Cloud |
| Version Control | Git & GitHub |
EduShield-AI/
│
├── app.py
├── requirements.txt
├── students_dropout.csv
├── EduShield_AI_Original_Notebook.ipynb
├── README.md
└── .gitignore
Clone the repository
git clone https://github.com/rinviriti/EduShield-AI.gitMove into the project
cd EduShield-AICreate a virtual environment
python -m venv .venvActivate it
source .venv/bin/activate.venv\Scripts\activateInstall dependencies
pip install -r requirements.txtRun the application
streamlit run app.pyDataset
│
▼
Data Cleaning & Preprocessing
│
▼
Exploratory Data Analysis
│
▼
Feature Engineering
│
▼
Model Training
│
▼
Performance Evaluation
│
▼
Explainable AI (SHAP)
│
▼
Interactive Streamlit Dashboard
EduShield AI can support:
- Universities
- Colleges
- Academic Advisors
- Educational Researchers
- Policy Makers
by identifying students who may benefit from early academic support and intervention.
- Deep Learning Models
- AutoML Integration
- Student Performance Monitoring
- PDF Report Generation
- Email Alert System
- Database Integration
- Real-time Student Analytics Dashboard
Rinvi Jaman Riti
Bachelor of Science in Computer Science & Engineering
Daffodil International University
GitHub: https://github.com/rinviriti
If you found this project useful, please consider giving it a ⭐ Star on GitHub.
Your support helps improve the project and makes it easier for others to discover it.
This project is released under the MIT License.