Machine Learning and Affordable Housing
CARL Director David Newton will be taking part in a panel discussion on affordable housing hosted by Omaha By Design on June 13th. One theme that will be discussed is the role that machine learning (ML) is currently playing and will play in the future in affordable housing. ML is a subfield of Artificial Intelligence dedicated to the development of algorithms that allow computers to learn how to accomplish tasks through pattern recognition and inference. Currently ML is being applied to affordable housing in three major ways:
As a means to make existing housing cheaper:
ML for city data analytics to forecast affordable housing stock.
Use of ML in a number of start-ups to make existing homes more affordable:
•Divvy helps consumers rent-to-own homes.
•Landed helps educators afford homes where they teach.
•Bungalow divides luxury apartments into multiple more-affordable units
As a way to improve urban planning and in turn make housing more affordable. Researchers are exploring how ML might be used to create flexible and dynamic zoning systems.
As a technology to streamline the design and building of homes to bring construction costs down.
Factory OS is a start-up that vertically integrates the entire residential building process to create pre-fabricated apartment buildings.
Advanced optimization workflows can leverage ML to make the design and building process more streamlined and optimized.
ML is also being used by investors and developers to identify areas of potential development and forecast housing trends. ML brings predictive power that previous technologies have not been able to muster. When combined with automated stock trading and investment, the potential for market speculation seems high. This may make housing more expensive. It’s also possible that other uses of ML by competing parties may help to balance this speculation out. It is very early days in the application of ML in these areas, so time will tell.