How can machine learning enhance the process of architectural design?

We explored the nature of machine learning and its potential applications to architectural design. Our exploration included identifying potential entry points for the applications and identifying design tasks that could most benefit from machine learning technology. We then developed two pilots that apply machine learning methods as a tool to improve the decision-making process: one focused on architectural visualization and the other on survey comments analysis.

Based on these pilots, we focused our efforts on the survey comments analysis because it proved to have the highest probability for design application. We utilized machine learning to create topic models that analyzed and graphically clustered comments from Gensler Workplace Performance Index surveys into defined themes, visually clustering the results around commonalities such as natural light and noise. We evaluated over 234,000 comments in a matter of seconds. We also explored opportunities to create targeted topic models and analyses by industries for more granular insights.