🎓 Student Performance Prediction System

An ML-powered web application that predicts student academic performance based on behavioral, academic, and engagement parameters — with an admin analytics dashboard, student search, and CSV export.

---

🔗 Live Demo

student-academic-performance-predict.streamlit.app

Login: admin / admin123 (admin) or staff / staff123 (staff)

---

📸 Screenshots

Prediction Form	Analytics Dashboard
Performance Distribution	Performance Heatmap
Student Records Table	Student Search

---

✨ Features

• 🔐 Role-based login — Admin and Staff access levels
• 🧾 Student data entry — study habits, assignments, participation, grades, sleep
• 🤖 ML prediction — classifies performance as Poor / Average / Good / Excellent
• 📊 Analytics dashboard (Admin only):
  • Total records, average absences, top result
  • Performance distribution bar chart
  • Study time vs absences heatmap
  • Student search functionality
  • CSV export
• 🔄 Multi-student support — form resets after each prediction
• 🧪 Automated UI testing — Playwright script to run 20 test cases automatically

---

🧠 ML Model

Detail	Info
Algorithm	Random Forest Classifier
Type	Multi-class Classification
Output Classes	Poor, Average, Good, Excellent
Test Accuracy	18/20 (90%) on automated test suite

Input features:

• Daily Study Time
• Study Consistency
• Assignment Completion Rate
• Class Participation
• Previous Semester Average
• Sleep Duration
• Age

---

🛠️ Tech Stack

Layer	Technology
UI	Streamlit
ML Model	Scikit-learn (Random Forest)
Data Processing	Pandas, NumPy
Visualization	Matplotlib, Seaborn
UI Testing	Playwright

---

⚡ Setup & Usage

1. Clone the repo

git clone https://github.com/rajit2004/student-performance-prediction.git
cd student-performance-prediction

2. Install dependencies

pip install -r requirements.txt

3. Run the app

streamlit run app.py

4. Login

Role	Username	Password
Admin	admin	admin123
Staff	staff	staff123

---

🧪 Automated UI Testing

This project includes a Playwright-based test script that automatically runs 20 student profiles through the app and validates predictions.

Setup

pip install playwright
playwright install chromium

Run tests

# Make sure the app is running first
streamlit run app.py

# Then in a separate terminal
python autofeed.py

Test results (latest run)

Passed: 18/20
❌ Sneha Sharma: Expected Average, Got Good     ← model edge case
❌ Pooja Krishnan: Expected Good, Got Excellent ← model edge case

The 2 failures are known edge cases where the model predicts higher than expected for borderline profiles — not bugs.

---

📁 Project Structure

student-performance-prediction/
├── app.py              # Main Streamlit application
├── autofeed.py         # Playwright automated UI test 
script
├── requirements.txt    # Dependencies
├── screenshots/        # App screenshots
└── README.md

---

🎯 Future Improvements

• Integration with real academic datasets
• Advanced ML models (XGBoost, Neural Networks)
• Feature importance visualization
• Student risk detection system
• Database integration (PostgreSQL / MongoDB)
• Automated daily test runs via GitHub Actions

---

💖 Support

If this project helped you, consider supporting my work:

---

👨‍💻 Author

Ranesh Rajit — B.Tech CS Student, India