Photo by Author
Over winter break I decided to use some of my free time to learn more about Tableau. It is a powerful tool that I’ve had some experience using for school projects, but I wanted to explore how it could be used to help answer high level questions about data. Given my interest in energy, I sought datasets that I could use to learn a bit more about renewable energy.
I came across the work Google is doing related to solar power called Project Sunroof [1]. I wondered if I could use their data to learn more about the diversity of investment in solar power across the US.
Based on my time living on both coasts and in the midwest I knew there were different perceptions surrounding solar power in each region. For example, it is a lot more common to see solar panels on a house here in northern California compared to where I grew up in Ohio.
While there are clearly differences in the weather patterns across the US that make solar power more viable in some places compared to others, I had to believe there was still potential for solar in places like where I grew up. Which brought me to the question I wanted to answer:
Which states have the biggest unutilized potential for solar power?
The rest of this article shows you how I obtained my results, hope you enjoy!
Project Sunroof is an undertaking by Google to make personal solar power more accessible. They’ve created a tool where users can enter their address and find out how much usable sunlight their house receives per year, how much of their roof is usable for solar panels, and their potential energy bill savings.
It also provides information on how to get solar panels, which helps residents understand steps they can personally take to transfer to clean energy.
The first step in the process was to login to the Google Cloud Console and go to the BigQuery tool. Google provides a large amount of data from Project Sunroof on the Google Cloud Platform BigQuery tool [2], which can be accessed using their free tier. Under the public datasets, Project Sunroof has two datasets, one grouped by postal code and one code by census tract. I decided it would be easiest to use postal codes to later summarize the data at the state level.
After querying all of the data, I downloaded it to a csv to be able to work locally. I created a database in PostgreSQL and imported the data. The dataset contained a large number of features and not all were necessary for the analysis. The SQL query below shows the information that was pulled from the complete set, grouped at the state level. Each feature has a comment describing the data being pulled in more detail.
With the data pulled, I was ready to create a visualization to start to make sense of the information. I used Tableau to create an interactive map of the US. The map colors each state according to the number of buildings that are available for solar power. The more buildings available, the deeper the orange color. Additionally, there is a bar chart showing the percent of buildings that have solar installed out of all buildings available for solar. For example, you see Hawaii at the top of the bar chart, representing that they have a high percentage of solar panels already installed on all of the buildings available for solar power.
Join Medium for free to get updates from this writer.
Subscribe
Users can filter by one or more states and click on either the state or corresponding bar to get access to more detailed information.