We are Potato Solar!

Members: Sergio Castellanos Brian Korgaonkar Ulhas Subramaniam Jonathan Maher

In here you will find our code for BERC's Cleanweb Hackathon 2016.

Inspiration 20-30% of U.S. residential solar energy system costs are directly attributed to lead-generation, a number that is far too high to continue the rapid uptake of solar and is the single biggest obstacle right now to the advancement of the solar industry. However, there is an effective way to reduce this figure and ensure more customers and installers succeed and install at the most economically efficient price possible. Big data can help to find high potential customers for solar, educate them with a no-pressure sale, match these customers to installers whose core-competencies are installation, not sales and marketing, and engage them to refer friends to go solar. By tapping this data, Potato Solar helps create a win-win situation all around: costs are reduced for all parties involved, the solar sale becomes quicker and easier, and solar penetration continues to grow.

What it does In the Potato Solar website, customers sign up with their address and facebook or linkedin or twitter logins, that's it. Using big data, Potato Solar then calculates their potential to buy, install and refer friends for solar using an algorithm continuously updated by machine-learning algorithms and expanding datasets. Potato Solar differentiates itself from other lead generation and targeting solutions in two key ways. First, by calculating the probability that a prospective customer can get financing - which is a crucial part of a solar sale that results in significant customer fall-off, and therefore wasted time, cost and effort, late in the sales cycle. And second, by scoring customers on their ability to provide referrals, which are by far the most cost-effective method for lead generation. All customers are provided competitive bids by various installers to ensure they get the best possible deal. For those customers with an extremely high probability of buying, Potato Solar will work hands-on with both the customer and installer to walk them through the entire install process and ensure that the customer gets the best deal possible. In the end, customers can continue to refer more of their friends and both can enjoy monetary benefits from referrals.

How we built it We used python, the Facebook SDK, the US Census Bureau API, and Project Sunroof data to determine the probability of customers buying and referring.

Challenges we ran into While working we needed to think through novel ways of downloading Project Sunroof data (as it did not have an API specifically), and to learn how to get the type of data we needed from Facebook and the US Census Bureau as these were all very different APIs. In addition, we did not have any front end engineers or designers, so we had to learn those skills ourselves, but we were happy to learn that.

Accomplishments that we're proud of We are proud that we developed a working code to download data from disparate sources such as Facebook, Project Sunroof, and the US Census Bureau while creating a simple and clean interface for inputs from our customers and our installers. The data we downloaded provides an effective measure against credit score and other hard to obtain data and our algorithm to develop a "probability of buying" will significantly reduce lead-generation costs for the installers we work with, thereby resulting in a more economical system for customers as well.

What's next for Potato Solar It is our explicit goal to turn this effort into a company. We believe that we are solving a real, tangible problem in the industry today, and have a novel approach that could potentially lead to a breakthrough. The steps we need to undertake include: gathering data and run a regression to ensure our probability of buying equation is as accurate as possible, and beginning to offer services to potential solar buyers and solar developers. It is our goal to build up the company to show a working prototype with at least 50 closed leads and then scale to help solar around the world.

Built With

Share this project:

Updates