An AI solution to help car manufacturers predict and reduce their cars' carbon footprint. By Aditya, Mathurah, Sanaa, and Muhammad.


After learning about the Volkswagen carbon emission scandal, we learned that data often about carbon emissions is inaccessible for the public. For car manufacturers, it's hard to predict and calculate carbon taxes and emissions for new models that are being developed.

What our solution does

Our solution is an artificial intelligence based application that allows the user to input the specifications of their car, specifically the engine size, cylinders, and combined fuel combustion to predict the carbon emissions of the vehicle, as well as classify the efficiency/eco-friendly level of the vehicle from 1-10 based on the emissions.

How it was created

The front-end/user interface was created with using the java swing library. Also the addition of Java awt library, incorporated many interact-able options and allowed to take input from the program. The use of external files allowed us to communicate between the 2 languages and acted as a bridge. The back-end uses a multi-linear regression ML model as the AI engine which trained to a near 90% accuracy. It was created with an array of python packages including pandas and matplotlib for data preprocessing/analytics, and Keras and Scikit-Learn to build and create the model. The trained model, after receiving and collecting user data from the front-end feeds it into the model and sends the predicted value back to the front-end.

Challenges we faced

We faced some main challenges in ensuring out AI model is accurate and also connecting the model to the inputted data on the Java application.

Accomplishments that we're proud of

We're proud that we have created a working demo in such little time, incorporating such exponential technology and skills required in artificial intelligence and data science. Additionally, we were able to train our model to about a 90% accuracy. We're also proud that we were able to solve an important problem in the world and target an issue concerning climate change.

What we learned

We learned important problem-solving skills, and how to leverage data and create models to produce accurate solutions.

What's next for KarboNinja

We're hoping to incorporate blockchain technology and put all the information on a ledger to make carbon emissions data accessible to consumers as well, for greater visibility.

Built With

  • python-java-html-css-pandas-matplotlib-keras-scikit-learn-machine-learning
Share this project: