Welcome to Munch - your go-to destination for smart, personalized restaurant recommendations! Whether you're gluten-free or just want to find the best burger in town, Munch will be your new foodie friend.
Access Munch here.
Video Pitch here.
Envitas is a dynamic project group passionate about bringing innovation to the food lovers community. Our team consists of:
- Jeffrey Taylor
- Topher Roebuck
- Gabriel Mendonca
- Cholyeon Cho
- Mike Yu
Munch is an innovative AI-driven project that revolutionizes the way users find and evaluate restaurants using advanced natural language processing techniques. With Munch, users can search for restaurants using detailed, conversational queries, and our AI system intelligently filters and ranks the results to provide the most relevant recommendations.
- Natural Language Search: Find restaurants with ease using intuitive, conversational queries.
- Detailed Review Analysis: Our AI has analyzed over 40,000 Yelp reviews from 4,400+ restaurants, scoring each review on key features like food quality, ambiance, and service to provide comprehensive evaluations.
- Aggregated Feature Scores: Review scores are combined to give you detailed ratings for 14,000+ unique restaurant features identified by our AI, offering nuanced insights beyond simplistic star ratings.
- Semantic Similarity Matching: Advanced text-to-SQL retrieval finds the most relevant restaurants based on the semantic meaning of your search.
- Informative Restaurant Cards: Easily view top recommendations, with overall ratings, location, and other key details presented in a user-friendly format.
Munch AI leverages a variety of cutting-edge technologies to deliver its advanced features and functionality:
- AI Frameworks & Libraries: LlamaIndex, GPT-3.5 Turbo API
- Backend: Flask, Python
- Frontend: HTML, CSS, JavaScript
- Databases: PostgreSQL
- Deployment: AWS (EC2, VPC)
- Version Control: Git, GitHub
- Containerization: Docker
- Natural Language Processing (NLP): Retrieval Augmented Generation (RAG), Sentiment Analysis, Text-to-SQL
- Data Processing: Pandas, NumPy
- Users input their restaurant search query in natural language on the Munch website.
- Our AI system analyzes the query and uses text-to-SQL techniques to retrieve the most relevant restaurants from our database based on semantic similarity.
- The retrieved restaurants are ranked and displayed to the user as restaurant cards, showcasing the overall ratings, location, and other important details.
- Users can explore the restaurant cards to gain insights into the specific features and ratings of each restaurant, helping them make informed decisions about where to dine.
To get started with Munch, simply access our website above and create an account. Once logged in, you'll have access to all of our AI helper and all the restaurant data we have available.
We welcome contributions from the community! If you have ideas for improvements or new features, feel free to submit a pull request.
If you have any questions or feedback, don't hesitate to reach out to us via this Github repository!