Machine Learning To Identify Why Detroit Residents Lack Access To Broadband

The University of Michigan’s Institute for Data Science (MIDAS) is conducting a year long, public-private research partnership with the City of Detroit and Microsoft to study where and why residents of the city lack access to broadband or devices. Microsoft is providing funding for the research.

The researchers will use machine-learning tools to determine what types of households in the city are likely to lack adequate access to the internet and/or devices, and their locations. Census data from 2019 has shown that more than a third of households in Detroit lack a broadband internet subscription.

MIDAS expects to be able to use the results of the study to make recommendations regarding which residents need better access to devices, as opposed to connectivity, and better understand how connectivity affects other quality-of-life aspects, such as health, education, and job placement.

“[City officials] want to know who the priorities are,” Jing Liu, the managing director at MIDAS, said. “They want to know what people need. For example, do people need devices? Do people have devices but they can’t afford a monthly plan to get internet access? Or can they afford both but they don’t know how to use the internet to effectively improve, for example, how they find health care providers, how they can search for jobs and so forth.”

MIDAS has already begun working with the city to collect broadband coverage and geospatial data. It is also an academic partner of Microsoft in its Airband initiative – a program which was launched in 2017 to use TV white space and wireless spectrum to connect rural communities to high-speed internet. Microsoft is now expanding the program to more metropolitan areas, and plans to use Detroit as a starting point.

“Partnering with MIDAS brings deep rigor to guide evidence-based policies, actions and learnings to address the digital divide in Detroit, grounded in an understanding of local context,” Vickie Robinson, the general manager of the Airband initiative at Microsoft, said. “These insights will be very valuable as we seek to increase access to affordable broadband, low-cost devices, and digital skilling resources in Detroit, and potentially establish replicable data tools that can be adapted for other places.”