Clean Markets showed the impact of clean energy by using machine learning to predict the ROI of green projects
When Clean Markets approached Promptworks, they were looking to build a tool to help demonstrate the ROI of green infrastructure to their clients. They wanted to help clients understand the benefits of improving buildings with green-focused projects like new HVAC systems or switching to fluorescent lighting.
They had developed a prototype data model and needed expertise and direction to turn it into a fully functioning web application that could spark interest in customers and consumers.
Here was our challenge:
Clean Markets helps utility companies and governments start and execute clean energy programs. They needed a tool to answer a common client need: demonstrate the ROI of integrating green infrastructure and emerging technologies into building and business improvements.
Internal development began on a proof of concept prototype called Building Energy Efficiency Retrofit Investment Model, or B-RIM. Quickly demonstrating its ability to calculate projections, Clean Markets looked for a partner that would take B-RIM from a spreadsheet full of formulas into an app.
B-RIM was good, but could it be better? The calculations were close, but for a commercially viable product, Clean Markets needed more accurate results for their clients. To make the case for new and emerging technologies, results needed to show a strong enough argument to convince large, slow moving institutions.
Here's how we solved it:
Create a web-based tool PromptWorks used a mix of off-the-shelf solutions and custom development to create a web-based platform. We used Ruby on Rails and Heroku to build a web application that allowed the Clean Markets team to build proposals. These proposals were powered by a custom tool that ingested specific details related to each property. Lastly, we created a Jupyter notebook that allowed Clean Markets to ingest their data and investigate different models.
This approach gave Clean Markets a tool their clients could use through a web-browser, removing the need to download a spreadsheet or install any software.
Provide more accurate results When our engineers dissected B-RIM, we realized there was a better way of building a statistical model. We utilized a linear model approach to create more accurate results than the rule-based approach in their proof of concept version and integrated machine learning to help improve and focus ROI predictions. The accuracy of these results proved to be a differentiator for Clean Markets and a conclusive factor in the success of the application.
See the results:
The new version of B-RIM became the basis of Clean Markets’ grant proposal for the Commercialization Challenge conducted by the Consortium for Building Energy Innovation, an organization composed of research universities, industrial firms and national laboratories dedicated to creating opportunities for energy reduction in existing buildings.
B-RIM won the Commercialization Challenge over five other finalists, out of an initial competition of 200 energy efficiency software technologies that were developed throughout the U.S. As the winner, Clean Markets has received grant proposal support, mentoring services, and networking opportunities from the CBEI’s Commercialization Center. Clean Markets is in the process of exploring additional funding opportunities to continue expanding the availability and application of B-RIM.
Interested in more examples of our work?
See how we helped ConnectDER offer grid oversight with a new dashboard or how Picwell helped consumers navigate health insurance using machine learning.
Contact us to discuss your next project.Contact