Inspiration - From our work in the industry, we know how much of a problem solar acquisition costs can be as well as local opposition to solar development. These are costly to the industry and we wanted to use data science to improve this process.
What it does - It allows a developer to screen areas that may be more open to solar development to avoid costly in person visits.
How we built it - Using a Machine learning algorithm based on data sets from NOAA, US Census, HUD, and OpenPV.
Challenges we ran into - finding the data and cleaning it up was a very time intensive process. We had to trade off potentially more interesting data in order to get a final product, due to time constraints.
Accomplishments that we're proud of - building a user facing interface and a solid data structure to build on in the future. Passes a "sniff test" and makes sense.
What we learned - Data is highly unavailable and developers need to be creative in data streams.
What's next for Project EDF - SolSource - Increasing the amount of features we can feed into the algorithm so that we can improve its predictive capability.