MIDS Capstone Project Spring 2020

Creative Cooking Club

Who does not know this situation? After a long day at work, you get home and scour the fridge for dinner. You see some old vegetables and think: “Not sure if I can cook something out of that. I guess I’ll settle for some pasta with tomato sauce”. This sounds sad, doesn’t it?

Besides monotonicity in diet and consequently a negative impact on health, this behavior is wasteful in terms of food and resources to produce them. This project aims to change that. We build an app, which suggests recipes to cook food that the user enjoys, enhances variety in diet, and makes use of ingredients already available.

The app is built in swift. It uses firebase and SQLite.

Three machine learning tasks are tackled: (1) User-recipe-matching based on preferences, (2) Image classification for ingredient detection, and (3) Ingredient swapping based on own trained embedding vectors and cosine similarity.

Last updated:

April 14, 2020