We wanted to track our carbon more accurately and with less margin for error, so we'd know how much carbon we need to offset/how much certain lifestyle changes would change our footprint.
What it does
It downloads your Monzo bank statement, then analyses each transaction to make a prediciton about it's carbon footprint. Then it tallys up your total footprint minus how much you've offset and displays that so you can use charities like https://www.atmosfair.de/de/ to offset the difference.
You can also correct it, when you do, it will update the prediction engine so in the future you and other users will have more accurate predictions.
How we built it
We build the front end using Svelte, and the backend using NodeJS, Express, and SQLite as our database.
Challenges we ran into
Coming up with the algorithms to combine the crowd sourced data in a way that gave users the feeling that they had total control over their data, while allowing other users to help make it more accurate.
What we learned
Don't do a massive redesign half way through an already tight deadline! ;P
What's next for Decim
Make the user interface responsive and maybe integrate it with a way to offset the carbon it's calculating, like maybe a crowd sourced wiki of charities with their effectiveness.