Inspiration:

What is the biggest existential threat we are facing as a society right now? Yep - climate change. And as students, we know just how lazy people are. So we went for the simplest, most effortless and probably most effective method - bringing carbon transparency to online shopping.

What it does:

It's a Chrome extension that displays the carbon output of products on Amazon. Once you add an item to your cart, your handy personal Greta Thunberg assistant will keep watch of it, let you know if you’re making a carbon-heavy purchase and help you reach your personal sustainability goals.

How we built it:

HTML, CSS and Javascript for the extension, and Python for an ML classifier

Challenges we ran into:

Chrome extensions have a lot of odd limits, like not being able to make API calls, so there was plenty to work around here. In general, transferring data and variables between different domains turned out to be surprisingly pretty tricky.

Accomplishments that we're proud of:

Submitting in the nick of time, and making a product that anyone can get working in a couple of minutes (just follow the README).

What we learned:

All of the odd restrictions and powerful abilities that Chrome extensions have - we can see why they're reviewed before being published!

What's next for Cotu:

Amazon has countless categories, and we can't build new hardcoded rules every time they add them. Therefore, we experimented used word embeddings and trained a neural network to map product types and descriptions to a fixed set of predetermined categories. This part is not yet integrated with the main model, and we hope to do so in future.

Built With

Share this project:

Updates