Inspiration

Without proper resources and experience, companies and businesses will have a difficult time evaluating their carbon emissions and effectively implementing initiatives to reduce their emissions. There are consulting firms such as BCG or Quantis that release their reports which present solutions that companies implement in an effort to reduce their carbon footprint. Along with this, information from certified environmental evaluations could be used in order to create a good estimate of a company's carbon footprint, which could be used in conjunction with the consulting data to suggest relevant ideas that would help offset company's carbon emissions. When we heard about Google's challenge of using AI and machine learning to help the climate crisis, we knew what we had to do.

What it does

Carbon Consulting allows businesses to upload PDFs that detail how they have affected the environment and receive feedback on how many carbon credits they have used and possible alternatives to some of their policies and actions. Companies will have to format their data in such a way that makes it easier for our interpreter to read. From the interpreted text, an AI will determine a category for the type of emission and related suggestions will be printed to the user.

How we built it

The front end was constructed using plain HTML and JavaScript. The back end uses Python and scikit in order to create a neural network and assign categories to sentence structures and key words.

Challenges we ran into

Time constraints involving implementing all the machine learning and data scraping components that we had wanted to.

Accomplishments that we're proud of

We successfully created a web server and a dataset to train our machine learning algorithm. We also were able to implement a plaintext analyzer to take in more abstract input.

What we learned

We learned a lot about Machine Learning, Artificial Intelligence, scikit, and how carbon credits are used and evaluated.

What's next for Carbon Consulting

Implementing data scraping and A.I. in order to automate the process of evaluating mission reports and generate suggestions based on public consultation reports. Eventually, we would like to be able to process financial statements in order to determine a company's true actions since mission reports can often be over-exaggerated.

Built With

Share this project:

Updates