Did you know that there is a relationship between the thermal mass of a building – the ability of a material to absorb and store heat energy – and how the building is used? For example, stone has a high thermal mass; it takes a lot of energy to make stone absorb heat. And how is the best way to use a building made of stone? The building’s efficiency rating goes up if you tend to use the building more frequently.

Watch how John Thomas of the IBM Competitive Project Office worked with Tooraj Arvajeh of green energy company BlocPower – using the IBM Data Science Experience notebooks to visualize the energy efficiency of various buildings in the Bronx – to run experiments with Watson services to bring together the different factors for each structure involving energy use, siting, building composition and design, heating systems and heat distribution systems, and other variables. The raw data and one-to-one comparisons are fascinating, but the engineers could not make any determinations from this veritable treasure trove of data.

Then Watson services steps in. It was able to provide a correlation, a bit of insight that let the engineers know that in actuality, they were really only dealing with three types of energy profiles: high (inefficient), medium, and low (efficient). The profiles it generated created a definition of each type that included the composition and design factors into the definition.

John explains:

“As a data scientist, now, with machine learning, when I have a new building to consider, even when it’s on the drawing board, I can add in the factors of the structure and get an energy efficiency score based on the buildings we’ve already measured.”

As a bonus, John goes on to demonstrate how to build that machine learning model.

More information