Recipository

Recipository logo

Inspiration

It's been a common experience of ours to find excess or odd grocery items piling up in our fridge. Like many, we also consider what recipe we could make out of this food. Yet, the human mind is nothing like that of a computer's when it comes to searching near-megabyte databases, it simply can't compare.

What it does

The goal of our project is to take a simple user input of a list of ingredient items, and return 5 great recipes with the most similar ingredients.

How we built it

Using the power of python's sickit-learn library, we were able to use Jaccard similarity to compare a user's input to thousands of recipes from an online database. Our main framework for the UI was good old fashioned HTML/CSS, with the goal of eventually putting these elements into react components for a truly dynamic website.

Challenges we ran into

Integrating the planned react frontend and flask (python) backend took a lot of our effort, and was ultimately outside of the scope of the project. In addition, multiple individual circumstances forced us to meet virtually and very infrequently, to the point that progress simply wasn't coordinated.