Electric Vehicle Range Prediction

Why this topic? As the world is moving towards cleaner resources of energy, electric vehicles have been at the center of this change. As the theme for this hackathon was AI for social good, we decided that this topic would be apt for this.

What does our model do? The main issue with electric cars is the uncertainty of the distance they will go in one charge. Customers are discouraged from buying electric vehicles due to this uncertainty and also the lack of charging stations. Our model takes features like battery capacity, average speed, tire type, city, environmental conditions, etc. and predicts the distance it will travel.

How did we build it? We gathered data for the Volkswagen e-golf model and collected various features from the car testing data. Then we used a variety of machine learning algorithms and processing techniques along with various visualizations to predict the distance the vehicle can travel.

Setbacks and obstacles we overcame: We had some trouble in finding the data for this topic. After finding the data we faced a lot of issues regarding it's preprocessing and cleaning. Initially the model accuracy was not up to the mark which was a setback which we later overcame.

Accomplishments that we're proud of: Our model for driving range produced an testing accuracy of 0.9 which is pretty good for a machine learning model.

What we learned from this experience: We learned various machine learning and data cleaning techniques throughout the hackathon. We also learned time management and teamwork during these 24 hours.

Future scope of this project: We can deploy it as an app where users can easily predict how much distance they can travel at any point of time. This can also help companies in deciding locations for charging stations so that charging stations are located at the right places We hope that this promotes the use of electric cars and it helps in betterment of the society by tackling problems like pollution and global warming.