🎓 Smart AI Academic Intelligence System

An advanced, full-stack predictive analytics and academic management platform built using the MERN Stack and Python Machine Learning models.

The platform enables educational institutions to proactively identify students at risk of academic failure or dropout, forecast future academic performance, and automate critical administrative workflows such as attendance tracking, internal marks management, and performance monitoring.

---

🚀 Features

• 📊 Student Failure Risk Prediction
• 🎯 Dropout Risk Assessment
• 📈 Next Semester CGPA Forecasting
• 👨‍🎓 Student Performance Analytics Dashboard
• 📝 Automated Attendance Management
• 📚 Internal Marks Distribution & Tracking
• 🔐 Role-Based Access Control (Student, Faculty, Admin)
• 📧 Automated Email Notifications
• 📂 Excel-Based Data Import & Processing
• 🤖 Machine Learning Powered Academic Intelligence
• 📜 Historical Prediction Logging & Analysis

---

🛠️ Tech Stack

🎨 Frontend

Technology	Purpose
React.js (Vite)	User Interface
Tailwind CSS	Styling
React Router DOM	Routing
React Context API	State Management
Axios	API Communication

---

⚙️ Backend

Technology	Purpose
Node.js	Runtime Environment
Express.js	REST API Framework
MongoDB	Database
Mongoose ODM	Database Modeling
JWT	Authentication
Bcrypt.js	Password Hashing
Node-cron	Scheduled Tasks
Multer	File Uploads
ExcelJS	Spreadsheet Processing
Nodemailer	Email Services

---

🤖 Machine Learning Layer

Technology	Purpose
Python 3.13	ML Runtime
Scikit-Learn	Model Training
Pandas	Data Processing
NumPy	Numerical Computation
python-shell	Node ↔ Python Communication
Pickle (.pkl)	Model Serialization

---

🧱 Project Architecture

The Smart AI Academic Intelligence System follows a Decoupled Client–Server–Machine Learning Architecture.

The React frontend communicates with the Express.js backend through REST APIs. The backend handles authentication, authorization, business logic, database operations, and ML orchestration. Predictive analytics tasks are delegated to Python-based ML services through the python-shell bridge.

                      [ User / Client Browser ]
                                │   ▲
                 HTTPS Requests │   │ JSON Responses
                                ▼   │
                      [ Express.js Server ]
                                │   ▲
         Mongoose Queries /     │   │ Data Payload
         Document Mapping       ▼   │
                     [ MongoDB Database ]
                                │
          ┌─────────────────────┴─────────────────────┐
          │                                           │
          ▼                                           ▼
   [ python-shell Bridge ]                 [ Training Pipeline ]
          │                                           │
          ▼                                           ▼
   [ predict_dropout.py ]                [ train_dropout_model.py ]
          │                                           │
          └────────────► [ Model Artifacts (.pkl) ] ◄─┘

---

🔄 Application Flow

1️⃣ Authentication & Authorization

• JWT-based authentication
• Route-level authentication middleware
• Role-based authorization
• Student, Faculty, and Admin access levels

2️⃣ Academic Operations

• Attendance recording
• Internal marks management
• Student profile tracking
• Department-level analytics

3️⃣ Machine Learning Prediction Flow

1. User requests prediction.
2. Express API receives request.
3. Prediction controller triggers Python service.
4. Python loads trained .pkl models.
5. Prediction is generated.
6. Result is returned as JSON.
7. MongoDB stores prediction history.
8. Dashboard displays analytical insights.

---

🗃️ Database Schema Architecture

User Schema

{
  name: String,
  email: String,
  password: String,
  role: ["student", "faculty", "admin"],
  createdAt: Date
}

Student Schema

{
  user: ObjectId,
  rollNumber: String,
  department: ObjectId,
  semester: Number,
  cgpa: Number,
  attendancePercentage: Number,
  academicStatus: ["Active", "At Risk", "Critical"]
}

Faculty Schema

{
  user: ObjectId,
  employeeId: String,
  department: ObjectId,
  subjectsTaught: [ObjectId]
}

Attendance Schema

{
  student: ObjectId,
  subject: ObjectId,
  date: Date,
  status: ["Present", "Absent"],
  recordedBy: ObjectId
}

Marks Schema

{
  student: ObjectId,
  subject: ObjectId,
  examType: ["Internal-1", "Internal-2", "End-Semester"],
  marksObtained: Number,
  maxMarks: Number
}

Prediction Schema

{
  student: ObjectId,
  dropoutRiskProbability: Number,
  predictedNextCgpa: Number,
  keyRiskFactors: [String],
  generatedAt: Date
}

---

🤖 Machine Learning Sub-System

train_dropout_model.py

Responsible for:

• Data preprocessing
• Feature engineering
• Dataset balancing
• Model training
• Model serialization

Outputs:

• smart_ai_dropout_risk_model.pkl
• smart_ai_performance_model.pkl

predict_dropout.py

Responsible for:

• Loading trained models
• Processing live student metrics
• Generating dropout probability
• Forecasting future CGPA
• Returning structured JSON results

---

⚙️ Local Installation & Setup

Prerequisites

• Node.js v18+
• MongoDB
• Python 3.10 – 3.13

Clone Repository

git clone <repository-url>
cd Smart-AI-Academic-Intelligence-System

Backend Setup

cd backend
npm install

Create .env

PORT=5000
MONGODB_URI=mongodb://localhost:27017/smart_academic_db
JWT_SECRET=YOUR_SECRET_KEY
EMAIL_USER=your-email@domain.com
EMAIL_PASS=your-app-password

Install Python Dependencies

pip install -r ml/requirements.txt

Seed Database & Train Models

node seed.js
npm run train:model

Start Backend

npm run dev

Frontend Setup

cd ../frontend
npm install

Create .env

VITE_API_URL=http://localhost:5000/api

Start Frontend

npm run dev

---

📈 Future Enhancements

• Deep Learning Models
• Student Recommendation Engine
• Parent Notification Portal
• LMS Integration
• Real-Time Risk Monitoring
• AI Academic Assistant

---

📄 License

This project is developed for educational, research, and institutional management purposes.