Odyssey

Inspiration

Our inspiration came from our collective frustration towards travel planning. While all of us have had separate experiences in planning trips, we all faced our own struggles, whether it be organizing activities to fit them all in a trip, or choosing events that satisfy all members in the trip. That is what we intend to fix.

What it does

Our app uses responses from a social media feed to train a model that predicts a travelers tastes and recommends events accordingly. It trains as the user gives feedback in terms of accepting or rejecting recommended events.

How we built it

We built it by combining multiple travel APIs, along with LLM processing that goes through posts and vectorizes them through our select tags. We then use that vectorized data to match events to users and adjust a traveler vector accordingly. We also have an android emulator that demonstrates our idea for the navigation aspect of our project.

Challenges we ran into

We ran into many difficulties, but out major issues were problems with LLM processing posts, and also having those processed posts becoming vectors. We also ran into a couple errors when trying to implement an IOS emulator, but we believe that overtime, we will be able to move past these errors.

Accomplishments that we're proud of

Personally, I'm proud of how well the team worked together and pulled through the multitude of trials. Collectively, we are proud of the design of our application and of the functionality of the demonstrable product.

What we learned

Over the course of 24 hours and multiple workshops, we learned a lot about training LLMs and using data produced by LLMs to train separate models. We also learned more about app design, and having a functional and efficient user experience.

What's next for Odyssey

Our next steps are cleaning up our current code, connecting all of our parts of the app, and implementing a live tracking map that will update events live.