The problem Decarbonate solves

Problem Statement
Increasing waste production from households and other activities leading to a higher carbon footprint.

Solution
To build an online platform to help people keep track of their daily carbon emissions caused by various activities and provide daily and monthly summaries of their carbon footprints.

USPs
Even though there are many carbon footprint calculators and social projects available online, we provide a system for storing daily carbon emissions, stored for a month, and then added up with other monthly carbon emission factors too, providing daily and monthly track of your carbon emission.
Along with that, our other major USP is the feature that allows users to just upload an image of a piece of waste, detect the class of that waste using deep learning and tell the carbon emission produced by the class of that waste material.

Our Goal
Making people realize and understand: As people get to know about how their carbon footprint is more than the average, they may feel the need to improve through small changes in their daily lifestyle if not by drastic ones.

Our Target Audience
Our major target audience includes people in the age group 18-45 years old, mostly college students to working professionals who are majorly responsible and even care for sustainable development in totality, while also hoping to attract more people.

Challenges we ran into

1. We used deep learning to classify images of waste items and find an average value of carbon footprint for it. We could not find a dataset suitable for our needs online and hence we had to build our dataset ourselves. Due to the short duration of the hackathon, we could not build a large enough dataset therefore the model could not perform well on the test set(new images).

2. We built our website's backend in Node JS but the image recognition model was built using python. Initially, we faced problems in integrating python and TensorFlow with Node JS. This problem was solved using the child spawn process functionality of Node JS where we could call a python script from our backend by providing the path to the python application.