💰 SpendSense — AI Expense Tracker

An AI-assisted expense tracking system with intelligent categorization and real-time financial insights.

SpendSense is a full-stack AI-assisted expense tracking application that helps users manage and analyze their finances intelligently. It combines machine learning with interactive dashboards to provide automatic expense categorization and meaningful insights into spending habits.

---

🚀 Live Demo

🌐 Frontend: SpendSense App
🔗 Backend API: API Endpoint

The application is fully deployed and accessible through the live demo links above.

---

📸 Screenshots

🔐 Authentication Page

📊 Dashboard

🧾 Expenses Page

📈 Insights

---

🛠️ Tech Stack

Frontend

• HTML
• CSS
• JavaScript
• React (CDN)
• Chart.js

Backend

• Python (Flask)
• REST API architecture
• JWT Authentication

Database

• PostgreSQL (Render — Production)
• SQLite (Local Development)

Machine Learning

• Scikit-learn
• TF-IDF Vectorizer
• Naive Bayes Classifier

---

✨ Features

• 🔐 User Authentication (Signup / Login)
• 💰 Add, Edit, Delete Expenses
• 🔎 Search & Filter Expenses
• 🤖 AI-assisted Expense Categorization
• 📊 Interactive Dashboard & Charts
• 📈 Spending Insights & Analytics
• ☁️ Fully Deployed (Frontend + Backend + Database)

---

🔌 REST API Overview

All endpoints require the header:

Authorization: Bearer <token>

except authentication routes.

---

🔐 Authentication

Method	Endpoint	Description
POST	/api/auth/signup	Register a new user
POST	/api/auth/login	Login and receive JWT token

Example signup request:

{
  "name": "Jane",
  "email": "jane@example.com",
  "password": "secret123"
}

Example login request:

{
  "email": "jane@example.com",
  "password": "secret123"
}

Example response:

{
  "token": "<jwt>",
  "user": {
    "id": 1,
    "name": "Jane",
    "email": "jane@example.com"
  }
}

---

💰 Expenses

Method	Endpoint	Description
GET	/api/expenses	List all expenses
POST	/api/expenses	Add a new expense
PUT	/api/expenses/<id>	Update an expense
DELETE	/api/expenses/<id>	Delete an expense
POST	/api/expenses/categorize	Get AI category suggestion
GET	/api/expenses/insights	Spending insights and analytics

Example expense request:

{
  "title": "Lunch at Zomato",
  "amount": 350.00,
  "category": "",
  "date": "2024-01-15",
  "notes": "Team lunch"
}

Example insights response:

{
  "total": 12500.00,
  "count": 45,
  "by_category": [
    {
      "category": "Food & Dining",
      "amount": 4200.00,
      "percent": 33.6
    }
  ],
  "monthly": [
    {
      "month": "2024-01",
      "amount": 6200.00
    }
  ]
}

---

🧠 AI Workflow

Expense Title → TF-IDF Vectorization → Naive Bayes Model → Category Prediction → Confidence Check

• If confidence ≥ 40% → ML prediction used
• If confidence < 40% → Rule-based fallback categorization

---

⚙️ Project Structure

spendsense/
│
├── backend/
│   ├── app.py                      # Flask application entry point
│   ├── requirements.txt            # Python dependencies
│   ├── expenses.db                 # SQLite database (auto-created on first run)
│   │
│   ├── models/
│   │   ├── __init__.py
│   │   └── models.py               # SQLAlchemy models (User, Expense)
│   │
│   ├── routes/
│   │   ├── __init__.py
│   │   ├── auth.py                 # Authentication APIs (signup, login)
│   │   └── expenses.py             # Expense CRUD operations & insights APIs
│   │
│   └── utils/
│       ├── categorizer.py          # Rule-based expense categorization logic
│       ├── train_model.py          # Script to train ML categorization model
│       └── model.pkl               # Trained machine learning model
│
├── frontend/
│   └── index.html                  # Single-page React interface (CDN-based)
│
├── screenshots/                    # UI screenshots for README
│   ├── auth.png
│   ├── dashboard.png
│   ├── expenses.png
│   └── insights.png
│
├── .gitignore                      # Ignored files (venv, cache, env files)
└── README.md                       # Project documentation

---

🔧 Setup Instructions (Local)

Follow these steps to run the project locally.

---

1️⃣ Clone the Repository

git clone https://github.com/kamalsharma001/spendsense.git
cd spendsense

---

2️⃣ Backend Setup

Navigate to the backend folder and create a virtual environment.

cd backend
python -m venv venv

Activate the virtual environment.

Windows

venv\Scripts\activate

Mac / Linux

source venv/bin/activate

Install dependencies and start the backend server.

pip install -r requirements.txt
python utils/train_model.py
python app.py

The backend will start at:

http://127.0.0.1:5000

---

3️⃣ Run the Frontend

Open a new terminal and run:

cd frontend
python -m http.server 3000

Then open your browser and go to:

http://localhost:3000

The application should now be running locally.

---

🚀 Deployment

• Frontend: Netlify
• Backend: Render
• Database: PostgreSQL (Render)

---

⚠️ Limitations

• Free database has limited lifetime on free tier
• Backend may experience cold start delay
• ML model accuracy depends on training dataset

---

🔮 Future Improvements

• 📱 Mobile application version
• 🔔 Budget alerts & notifications
• 📤 Export reports (PDF / Excel)
• 🧠 Advanced ML categorization models
• 👥 Multi-user financial analytics

---

👨‍💻 Author

Kamal Sharma
B.E. Computer Science Engineering (AI & ML)
Chandigarh University

📧 sharmakamal1210@gmail.com
🌐 GitHub: kamalsharma001