The automotive industry is undergoing an electric revolution, as the world embraces the next generation of sustainable transportation. However, the availability of reliable electrical infrastructure is a major constraint for mainstream EV adoption. As more private and public investment is made into tackling this problem, it is crucial to start thinking more about questions such as the optimization of resources.

A team of UC Berkeley Data Science students, curious about electric cars and the future of the automotive industry, decided to tackle this very problem by creating a POC data analysis project of EV adoption and infrastructure sustainability based on data from Washington State. The team performed exploratory data analysis to figure out what variables were indicative of EV adoption rates year over year and used this data to train several statistical forecasting models, including linear regression and ARIMA. The team then combined the output of these machine learning models with electric generation and consumption data, granular to the ZIP code, to determine which ZIP codes in Washington State lack the electrical infrastructure to support the projected influx of the EV population. Finally, the team generated a user interface displaying a map of which ZIP codes have electrical deficits.

How the electric revolution unfolds in the coming decades is still largely dependent on government regulation and corporate investment, but one thing is for sure: it is inevitable. How policy frameworks and incentive mechanisms for customers are created is largely up to public entities, but this team showed how they could use their skills to drive policy in the right direction and encourage public awareness.

Built by: John Sha, Lawrence Chen, Pancham Yadav, Rohan Srivastava, Tina Ye

Posted in Automotive, Sustainability, Transportation and tagged , , ,