Food Party

Inspiration

Finding something to eat that you'll love is difficult. Doing that with a group of people? Impossible.

Until now. Food Party aims to eliminate this problem in the easiest way possible.

What it does

Users go through a quick onboarding where they state their allergies and generic food preferences (veggie, vegan, etc.)

Then the fun begins: We show you just 5 recipes. You tell us which ones you like or don't. We then create a unique profile that recommends you the perfect recipes based on this simple selection.

Where's the party?

The outstanding feature is that this not only works with you, but also with your friends, family, coworkers and pets! (Ok, maybe not that last one)

How we built it

Our frontend was designed with Figma and built using Vue. Our backend consists of a simple Python Flask app and a database. The heart of our recommendation engine lies in a pre-trained recommender model (using k-NN) that enables collabaritive filtering. It's a bit of math, but basically works like this: "User X liked A and B. User Y liked A. So Y will probably also like B" (but a bit more complicated of course)

Challenges we ran into

Time pressure makes things difficult and frontend is harder than we thought. (Databases are also pain)

Accomplishments that we're proud of

The idea of our recommender model. There is no need of large transformer model or having to understand what each ingredient of a recipe actually does. KISS and do it like Spotify, Netflix and other giants.

We are also in love with our UI design. See also the Figma link 👀