E-commerce Recommendation System

Overview

This project is an e-commerce platform with a personalized product recommendation system. The application uses content-based filtering to suggest products to users based on their search queries and interests.

Features

• Content-based product recommendations
• User authentication (sign-up and sign-in)
• Trending products display
• Responsive design using Bootstrap
• Theme customization (default, dark, green)
• Display zoom functionality

Tech Stack

• Frontend: HTML, CSS, JavaScript, Bootstrap 4
• Backend: Flask (Python)
• Database: MySQL
• Machine Learning: scikit-learn for TF-IDF vectorization and cosine similarity calculation

Prerequisites

• Python 3.7+
• MySQL
• Required Python packages (listed in requirements.txt)

Installation

1. Clone the repository:

git clone <repository-url> cd ecommerce-recommendation-system

2. Create and activate a virtual environment (optional but recommended):

python -m venv venv

source venv/bin/activate # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Set up the MySQL database:

• Create a database named 'ecom'
• Update the database connection string in app.py if necessary:

app.config['SQLALCHEMY_DATABASE_URI'] = "mysql://root:2906@127.0.0.1:3306/ecom"

5. Initialize the database tables:

• Add these lines to app.py before the if name == 'main' block:

with app.app_context(): db.create_all()

6. Update file paths:

• Modify the file paths in app.py to match your dataset locations:

trending_products = pd.read_csv("/path/to/your/trending_products.csv") train_data = pd.read_csv("/path/to/your/clean_data.csv")

Running the Application

1. Run the Flask application:

python app.py

2. Open your web browser and navigate to:

http://127.0.0.1:5000/

Project Structure

app.py: Main Flask application file templates/: HTML templates

index.html: Landing page with trending products main.html: Main page with search functionality

static/: Static files (images, CSS, JavaScript)

How the Recommendation System Works

The recommendation system uses content-based filtering with the following approach:

1. TF-IDF vectorization is applied to product tags/descriptions
2. Cosine similarity is calculated between the query product and all other products
3. Products with the highest similarity scores are recommended to the user

Future Improvements

• Implement collaborative filtering
• Add user browsing history for better recommendations
• Integrate product categories and filters
• Add a shopping cart functionality
• Implement session management
• Enhance UI/UX design

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.