Restaurant Recommendation System

---

Project Description

• This project focuses on developing a smart restaurant recommendation system that delivers personalized suggestions based on user preferences, budget, dietary restrictions, and current location.
• A hybrid recommendation model combining collaborative filtering and content-based filtering techniques is implemented to provide accurate and relevant results.
• The system adapts dynamically based on user inputs and geolocation data, enhancing the user's decision-making process when exploring dining options.

Screenshots

Home Page:

Input Page:

Recommendation Output:

Installation and Setup

Using Conda (Recommended)

# Create a new conda environment
conda create -n restaurant_recommender python=3.10

# Activate the environment
conda activate restaurant_recommender

# Clone the repository
git clone https://github.com/sagarb1511/Datascience.git
cd Restaurant-Recommendation-System/Flask

# Install dependencies
pip install -r requirements.txt

Using Python venv

# Create a virtual environment
python -m venv venv

# Activate the environment
# On Windows:
venv\Scripts\activate
# On macOS/Linux:
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

Running the Application

# Navigate to the Flask directory
cd Flask

# Run the Flask web application
python app1.py

# Visit http://localhost:5000 to access the app

Project Structure:

Restaurant-Recommendation-System/
├── Documentation...
├── Flask/
│   ├── __pycache__/                # Compiled Python cache files
│   ├── static/                     # Static assets (Images)
│   ├── templates/                  # HTML templates for the frontend
│   ├── app1.py                     # Flask application entry point
│   ├── Final_Development_Phase.ipynb  # Flask dev notebook
│   ├── requirements.txt            # Python dependencies
│   └── restaurant1.csv             # Restaurant dataset
├── Model/
    └── Final_Development_Phase.ipynb  # Model training and evaluation notebook

Technologies Used

• Python 3.10 – Core programming language
• Flask – Lightweight web framework to build the backend API
• Scikit-learn – For building and evaluating machine learning models
• Surprise – Specialized library for collaborative filtering (e.g., SVD)
• Pandas / NumPy – For data manipulation and preprocessing
• NLTK – For natural language processing and cleaning review text
• Matplotlib / Seaborn / Plotly – For data visualization and EDA
• HTML / CSS / JavaScript – For designing the frontend interface

---

Model Architecture

This project uses a Hybrid Recommendation Model combining the strengths of:

🔹 Content-Based Filtering

• Analyzes restaurant attributes like cuisine type, average cost, rating, and delivery option.
• Matches these with user-stated preferences to recommend relevant restaurants.

🔹 Collaborative Filtering (SVD)

• Uses historical user rating data to find similar users and suggest restaurants based on collective behavior.
• Implemented using the Surprise library's Singular Value Decomposition (SVD) algorithm.

🔹 Hybrid Approach

• Merges both filtering strategies to solve the cold-start and sparsity problems.
• Balances personalization with discovery of new or less popular options.

Dataset

The dataset is sourced from Kaggle and titled:

Zomato Bangalore Restaurants Dataset by Himanshu Poddar

Conclusion

This project successfully demonstrates how machine learning and recommendation systems can simplify dining decisions by tailoring restaurant suggestions to user preferences and behavior. It offers a powerful and adaptive solution for users in both familiar and unfamiliar areas. The hybrid model ensures balance between personalized results and discovery of new options. This project lays the foundation for more advanced, real-time, and location-aware food recommendation engines.