Restaurant-Tinder

• 

  Frontend look like

Inspiration: We were inspired by the everyday struggle people face when choosing a restaurant. Searching online often yields overwhelming results. Our goal is to make restaurant selection easy and fun!

What it does: Users see nearby restaurants and select 5 to start. The app then provides suggestions, and users simply click 'like' or 'dislike' to tailor the results.

How we built it: The app is consisted of frontend written using vue and the backend using pyton flask. I used primevue library for UI components. The 2 parts communicate with each other via a rest api. Our backend utilizes data stored in Firebase. Initially, we employ K-Prototypes clustering to categorize and label all restaurants. When a user selects restaurants they like, our system analyzes the reviews and labels associated with those specific choices. This analysis feeds into our proprietary scoring system, which calculates a final score to generate personalized restaurant suggestions.

Challenges we ran into: Making the picture and restaurants information shape changing while the website change took a lot of time. But this technology can provide user use it when they open it on phone. How to make the suggestions for user, cause the data is full of categorical data and the reviews for each restaurant have only five. So designing a way to score restaurants and actually implementing it took some time.

Accomplishments that we're proud of: Completing this project was important to us because we genuinely wanted to solve the common struggle of choosing a restaurant, especially since neither of us was initially familiar with frontend development. We are particularly proud of designing the scoring algorithm and implementing the backend logic. This involved using Machine Learning (ML) and cosine similarity to analyze reviews, assigning appropriate weights, and combining these factors together to generate meaningful recommendations

What we learned: We learned a lot of frontend, and how to actually implement ML to actually world problem. Connect to firebase and upload the data on it.

What's next for Restaurant-Tinder: This project has clear potential to be scaled and optimized into a business venture, given how much people struggle with choosing where to eat. We want to put the code onto google cloud service.

Built With

• firebase
• flask
• google-icon
• google-maps
• python
• tailwind
• vue