We are bunch of tree-hugging eco-warriors who happen to also be interested in AI. Our world is going in the wrong direction, but our app will single-handedly save it!
What it does
Given a food recipe, eco-friendly substitutions are recommended with results dependent on location and time of year.
How I built it
From a food dataset of over 1 million recipes, language models were trained to learn semantic relationships between ingredients of the different recipes. We have built models that can take an ingredient or recipe as input and outputs the best alternative ingredients that have lower carbon footprints with the seasonality of ingredients and location of the person taken into account.
This was built as backend code for a website built with azure and flask that can take a user input recipe or a photograph of a recipe, in which case the text is converted using azures text to image software, and will return the best substitutions one can make to be more eco-friendly.
Challenges I ran into
Accomplishments that I'm proud of
We fixed 4,782,395 bugs.
What I learned
How to make a website and deploy it. Connection of different technologies into a single framework.
What's next for PlanECO