Restaurant Recommendation System

Project Overview

This project is a FastAPI-based restaurant recommendation system powered by a content-based filtering model. Users can search for a restaurant and refine recommendations using budget, rating, and cuisine preferences. The project also includes data visualizations that summarize restaurant trends from the dataset.

Features

• Restaurant recommendations based on similarity
• Filtered results using budget, minimum rating, and cuisine
• Friendly error messages when a restaurant is not found or no filtered matches are available
• Login and registration with SQLite-backed user storage
• Insights page for dataset visualizations
• Popular cuisines section on the homepage

Tech Stack

• Python
• FastAPI
• Scikit-learn
• Pickle
• HTML
• CSS
• Jupyter Notebook

Screenshots

Home Page

Insights Page

Dataset Visualizations

How to Run

1. Create and activate a virtual environment.
2. Install the required packages.
3. Run the FastAPI app from the project root.

pip install -r requirements.txt
uvicorn app:app --reload

4. Add your default login credentials in .env:

APP_USER_EMAIL=shreeyash2573@gmail.com
APP_USER_PASSWORD=Shreeyash@2516
FLASK_SECRET_KEY=restaurant-recommender-secret-key

5. Open the local FastAPI URL in your browser, usually http://127.0.0.1:8000.
6. Log in with the seeded user or create a new account from the register page.
7. Visit /insights to view the saved visualizations in the web app.

Project Structure

app/
  static/
    css/
    images/
  templates/
    index.html
    result.html
    insights.html
app.py
model/
  build_model.ipynb
  recommend.py
  restaurant.pkl
  tfidf.pkl
  tfidf_matrix.pkl

Notes

• Visualization assets are served from app/static/images/.
• The notebook remains available for model building and analysis.
• The web app now exposes the charts through the /insights route.