Inspiration
We care for the environment. The evidence that we are ruining our beloved home is overwhelming. We want to try and spread the message of harmful carbon emissions and how much your car really flows out of its exhaust pipe. It's much more than you would think. We want to inform our peers about this pressing issue.
What does it do
It is a trained Neural Network on a data set of cars and their specifications along with their CO2 produced per kilometer in grams. An app front end that can be built onto Android, IOS, OSX, Windows, or Linux, uses this Neural Network to take your car's specs and give you the amount of grams per kilometer of CO2 your car likely produces.
How we did it
We used a dataset from online https://old.datahub.io/dataset/car-fuel-consumptions-and-emissions to train a Artificial Neural Network using the PyBrain library. We saved this network into an .xml file and used it into a kivy interface for the user to input values and get their result back
Challenges we ran into
We tried to not use a library to do the machine learning, and it turns out that's a bad idea. As well, there was plenty of problems with overfitting, formatting issues with the dataset, and all around hours of frustration. But, in the end, we overcame and finished the App.
What we are proud of
We are most proud that the code actually works, and that we have a presentation and a cool logo to boot.
What we learned
To use a library and not to do it yourself
Log in or sign up for Devpost to join the conversation.