Fynder

Fynder is like Tinder for experiences—enter a location, swipe through nearby attractions, save what excites you with right, skip with left, and discover the best things to do around you.

Comment

Inspiration

After scouring through Yelp and Tripadvisor for summer vacation ideas, we realized that the tools available for proper research for a great time just did not exist. That's why we built Fynder, an app that learns based on your preferences to recommend fun things to do near you that you genuinely care about.

What it does

Fynder recommends activities, attractions, and experiences tailored to your personal tastes. It considers your past choices, preferences, and even contextual factors like location and time of day. Users can interact with a chatbot to refine recommendations and discover hidden gems they might not have found otherwise.

How we built it

We built a full-stack web application with a frontend written with HTML, CSS, and JavaScript. The backend was written in Node.js with a Python integration. We created a DQN machine learning model using tensorflow in python to give every user personalized recommendations based on their previous swips. This model is able to learn with the user to prioritize the amount of liked places. We also integrated superbase to a chatbot integrated in the site so that it is able to answer user questions using the information we have on locations in our own database. This lets the chatbot better answer the user's questions by knowing what they are looking at and talking about.

Challenges we ran into

We ran into many challenges as we went along, but among some of the most difficult to solve were integrating our python modules into node js and setting up the chatbot with supermemory.

Accomplishments that we're proud of

We successfully built an end-to-end system that learns from user interactions and improves over time. Fynder currently supports 300+ attractions, provides personalized recommendations in real-time, and maintains a chatbot that can remember user choices. The system effectively combines AI with user-friendly design, which was a major achievement for a hackathon project.

What we learned

We learned about how important it is to split up work while working as a team and how if there is a feature that only 1 person is working on but that all others need it to continue, it can lead to project inefficiencies.

What's next for Fynder

We hope to expand Fynder's memory capabilities and add more data to our database. So far, Fynder supports 300+ attractions, but we want to capture every choice of fun for every type of individual. We also want Fynder to be able to book tickets, consider fun things to do on long trips to attractions, and be the go-to place for itinerary making for any activity and party size, small or big.

Share this project:

Updates