DeepFashion - E-commerce Platform with AI Recommendations

Project Overview

DeepFashion is a full-stack e-commerce web application specializing in fashion retail. The platform features a complete shopping experience with user authentication, product catalog, shopping cart, order management, and an AI-powered recommendation system that provides personalized product suggestions based on user behavior.

Technology Stack

Frontend

• React.js - User interface framework
• Material-UI - UI component library
• React Router - Client-side routing
• React Context API - State management

Backend

• Node.js - Runtime environment
• Express.js - Web application framework
• MongoDB - NoSQL database
• Mongoose - MongoDB object modeling
• JWT - Authentication
• Imgur API - Image hosting

AI/ML

• Python - Machine learning implementation
• Implicit - Collaborative filtering library (ALS algorithm)
• Pandas - Data processing
• NumPy - Numerical computations

Admin Panel

• React.js - Admin interface
• Vite - Build tool

Project Structure

DeepFashion/
├── frontend/          # React.js client application
├── backend/           # Node.js/Express API server
├── admin/             # Admin panel for product management
├── AI/                # Machine learning recommendation system
├── data/              # Product and comment data
├── new-data/          # Additional data resources
└── utils/             # Data processing utilities

Installation & Setup

Prerequisites

• Node.js (v14 or higher)
• Python 3.8+
• MongoDB (local or Atlas cloud)
• npm or yarn

Backend Setup

cd backend
npm install
# Create .env file with:
# MONGO_URI=your_mongodb_connection_string
# IMGUR_CLIENT_ID=your_imgur_client_id
npm start

Frontend Setup

cd frontend
npm install
npm start
# Runs on http://localhost:3000

Admin Panel Setup

cd admin
npm install
npm run dev
# Runs on http://localhost:5173

AI Recommendation System

cd AI
pip install pymongo implicit pandas numpy scipy schedule
python main.py
# Continuously runs recommendation pipeline

Features

User Features

• User registration and authentication
• Product browsing by category (Men, Women, Kids)
• Product search functionality
• Shopping cart management
• Order placement and tracking
• Product reviews and ratings
• Personalized product recommendations
• User profile management

Admin Features

• Product management (add, edit, delete)
• User management
• Order management
• Dashboard with statistics

AI Features

• Collaborative filtering recommendation system
• Real-time recommendation generation
• User behavior analysis
• Product recommendation API

API Endpoints

Products

• GET /product/allproducts - Get all products
• GET /product/:id - Get product by ID
• POST /product/addproduct - Add new product (admin)
• GET /product/category/:category - Get products by category

Users

• POST /user/signup - User registration
• POST /user/login - User login
• GET /user/:id - Get user information

Cart

• GET /cart/:userId - Get user's cart
• POST /cart/addtocart - Add item to cart
• DELETE /cart/removefromcart - Remove item from cart

Orders

• POST /order/addorder - Place new order
• GET /order/allorders - Get all orders (admin)
• GET /order/:userId - Get user's orders

Recommendations

• POST /recommendation/getrecommendations - Get personalized recommendations
• GET /recommendation/allrecommendations - Get all recommendations (admin)

Development Notes

Data Processing

The utils/ folder contains scripts for:

• Data preprocessing and normalization
• Product data upload
• Comment data generation
• Transaction extraction
• Image downloading

AI Model

The recommendation system uses Alternating Least Squares (ALS) collaborative filtering:

• Trained on user transaction history
• Generates top-N recommendations per user
• Updates periodically via scheduled pipeline
• Stores recommendations in MongoDB

Deployment

The application is designed to run on localhost. For production deployment:

1. Set up environment variables
2. Configure MongoDB Atlas connection
3. Set up image hosting (Imgur or alternative)
4. Build frontend: npm run build
5. Deploy backend to cloud service (Heroku, AWS, etc.)

Project History

This project was originally developed as a group project for a Software Engineering course (SE4AI) completed in June 2024. The project has been moved to this personal repository to serve as a portfolio piece. Future improvements are planned but not yet implemented.

---

Acknowledgments

Original Collaboration (Course Project - Completed June 2024)

This project was initially developed as a collaborative effort with the following team members:

Team Members:

• Huy (myself) - Backend architecture, AI recommendation system, frontend components (Cart, Order, Product pages, Comments, Recommendations), payment system, data processing
• Cường - Database design, Admin panel development, backend optimizations, UI improvements
• Trí - Frontend development, UI/UX design, Header/Footer components
• Hân - UI/UX design, frontend components, testing
• Toàn - AI model research, data preparation, frontend components

Key Collaborative Contributions:

• Database schema design and implementation (Cường)
• Initial frontend structure and routing (Trí, Hân)
• UI/UX design and Figma mockups (Hân, Trí)
• Admin panel core functionality (Cường)
• AI model selection and initial implementation (Toàn, Hân)
• Code cleanup and optimization (Cường, Hân)

Project Timeline:

• Course start: March 2024
• Final presentation: June 28, 2024
• Course completion: June 29, 2024