Interview with Angela Bassa, Data Science Manager at EnerNOC

July 2016

Angela Bassa is in charge of the data science team at Energy software intelligence provider EnerNOC — ultimately helping over 1,100 software subscription customers serving over US$1 billion in customer savings to date. With an impressive career trajectory discovering data science while studying math at MIT, Angela also serves as Data Science Technical Advisor to Boston-based MedTech startup Mirah focusing on making mental healthcare more objective and data-driven. We find out how about her work at EnerNOC.

WTB: In the energy sector how do you identify which is useful data?

In our job, data is everything. It’s the cacao bean to our chocolate! EnerNOC is involved in the data business from ground up, using our own energy data collection devices and APIs, forging unique partnerships with utilities and solar providers, running our own state-of-the-art Network Operations Centers (the NOC in EnerNOC), and sending real-time alerts to our customers with predictive recommendations on how to manage their energy.

The challenge is making sense out of the unbelievable complexity that is the energy sector. Our data is made up of weather models, consumption tracking, tariff structures, billing details, geographic nuances, local and global regulations, sector and industry benchmarks, macro trends, and much more. The data science team is responsible for turning all of that “messy data” into information that is useful, valuable, and actionable for each and every customer, every day.

WTB: Once actionable, how do clients use the data?

The best example I can think of is the implementation of the patent-pending novel weather-normalization feature we have in our energy intelligence software (EIS) allowing like-for-like comparison of energy consumption for different time periods with differing weather conditions, enabling you to track a facility’s relative performance over time without worrying about the impact of varying outside air temperatures.

E.g., what if it’s particularly hot during the second week of August and our facility manager wants to see whether her building is operating inefficiently or just meeting demand for more air conditioning? With one button, she can turn the weather normalization feature on in our application and see that when the effects of temperature are removed, her building actually consumed 5% less energy during the hotter week than it did during the cooler week (see graph), meaning that her building was performing optimally and simply using more energy because it was exceptionally warm.

Communicating the complexity of weather impacts on energy consumption with a simple and intuitive visual is incredibly valuable to our customers, and in turn to our company.

WTB: Which Computing platform do you use?

We have invested more than US $200 million in technology to date. Our data pipelines are built using Kinesis queues and Lambda processing units and our core interval database utilizes both DynamoDB and Redshift to power our feature-rich interactive analytics applications, which are built using AngularJS.

We are also re-building our NoSQL-based time-series analytics platform with the latest server-less architecture technologies available on AWS. Other services can deliver custom datasets to the analytics platform through our data lake, and deep analytics are performed in parallel using Spark and R, leveraging the built-in scalable nature of the server-less architecture. Our entire system is built as a set of micro services, whose API’s are published and discoverable through a central services catalog.

WTB: What advice would you give to budding AI developers?

The pace at which our field keeps growing is wonderful and amazing and a testament to how much can be achieved, but it is also seductively distracting. It is easy to become enamoured with the latest trend or tool, and the more complex models may not be the most appropriate for every application. My advice would be to start simple: break a problem down to its atomic components, then build it back up towards a solution, and only ever add complexity if a simpler model doesn’t get the job done.

WTB: What’s the most exciting part of your job?

When I come into the office, the elevator doors open up on the 5th floor at One Marina Park to a beautiful bright wall showcasing the granted patents and applications that EnerNOC has filed to date (check it out here). When I started, my goal was to get someone on our team featured on that wall. I see dozens of patents now, and almost a third of them are from our team! It’s a constant reminder of the importance, value, and innovativeness of the remarkable people I work with, and it fills me with pride.

Additionally the speed of innovation in the field is vertiginous, and that’s really a function of how young the discipline really is. I’m excited to see what the next chapter in data science brings.

WTB: Are you excited about speaking at AI With The Best?

I am very excited to participate in the AI With The Best conference. I love the concept of connecting with hundreds of developers from around the world who are looking to share knowledge and may not otherwise be able to attend a conference of this caliber in person. It’s a brilliant project, and I can’t wait to for it!

Thanks so much Angela for chatting with us!

Find out more about “Machine Learning in Production: integrating with the software stack” with Angela 2pm Saturday 24th EDT at AI With The Best online conference September 24–25th 2016