Foodprint

Web application that allows user to visualize their carbon emission by date and food categories
Proposed mobile application for user to keep track of their carbon emission
Entering recipe or uploading receipts allow the user to view the list of the food filtered by Cohere.ai NLP API before proceeding

Inspiration

Climate change is driven by endless consumerism, while it is practically impossible to stop this, individual habits can go a long way. Many people don't know that the emissions of beef are about 50x higher than chicken, and chicken is already higher than most vegetables and high protein alternatives. Providing an understanding of these issues can allow consumers to make better choices for the environment.

What it does

Processes grocery receipts and recipes and matches the individual items to their respective carbon emissions rating. Provides users with data on emission history and emission compared to other users.

How we built it

Uses React and Javascript for the front end and python for the backend and communication is done through flask. The application uses a combination of google vision API and cohere’s NLP to process images of grocery receipts. Uses CockroachDB to store grocery emission data of specific items, as well as user specific emission data.

Challenges we ran into

As many of us were first time users of react we ran into a few problems with the frontend. Other than that, connecting individual work was the biggest challenge but we generally got it together. Additionally this hackathon was far more time constrained than others, which is a challenge we had to go through.

Accomplishments that we're proud of

Worked in a team to put out a prototype and base web framework
Actually going to bed

What we learned

Learned how to communicate between the frontend and backend using flask
Use of CockroachDB and SQL queries.
Tuning using cohere

What's next for Foodprint

We will be finishing up the web application we have been working on during the hackathon, improving the backend and frontend. We will also be enhancing the NLP model in Cohere AI to get a more accurate filtering from the texts on the receipts. On top of the web application, we plan to complete the mobile apps in iOS and Android.