UI Professor develops model to predict wireless use, help create more effective public policy

2/13/2013 Kim Gudeman

When University of Illinois professor Christian Sandvig looks at his map of South Central L.A., he sees red. The same goes for a small town in the Amish country of Illinois and Chicago's Lincoln Park.

Written by Kim Gudeman

When University of Illinois professor Christian Sandvig looks at his map of South Central L.A., he sees red. The same goes for a small town in the Amish country of Illinois and Chicago's Lincoln Park.

Prof. Christian Sandvig has developed a new model to predict wireless density.
Prof. Christian Sandvig has developed a new model to predict wireless density.
Prof. Christian Sandvig has developed a new model to predict wireless density.

Sandvig’s map is part of his research, The Red Project, a technique that predicts Wi-Fi density per city block. As more and more wireless technologies can be bought by anyone, policymakers have struggled to understand the implications of wireless systems that aren't controlled centrally like a radio or television station. Not only could too much traffic on wireless systems in the same area lead to slower Web surfing, with enough users packed closely together, wireless products from bluetooth headsets to baby monitors might stop working.

His work was recently published in the journal Vectors, issued by the University of Southern California, and in the proceedings of the Annual Research Conference on Communication, Information, and Internet Policy.

“We now have this incredible technology that allows us to access the Internet in places where we never thought we could without a wire,” said Sandvig, an associate professor of communications and researcher in the Coordinated Science Laboratory, who studies communications infrastructure and policymaking. “These days, the challenge has become figuring out how to organize the infrastructure so that it continues to serve the population.”

Unlike similar models, which make predictions about wireless use in the future based on population density or the presence of radio signals right now, Sandvig’s model incorporates income, age, race and other data pulled from the U.S. Census into the equation.

The model is complex: While lower-income neighborhoods are less likely to have dense wireless activity, other factors may complicate the issue. College students, who are among the lowest income earners, are very likely to be wired. And conditions change block-by-block: some low-income areas like South Central L.A. may have pockets of very high wireless use.

Likewise, high-income neighborhoods tend towards high wireless density. But a community of high-income seniors may be less wired than a neighborhood of middle-class workers, who see the technology as a way to advance their careers and income.

“We’ve taken what we know about people and what we know about wireless technology and created a model that can give a pretty accurate view of wireless density in a given area,” Sandvig said.

His research team verified the model by cruising neighborhoods and reading wireless signals with an antenna. The team tested the model in Los Angeles, Chicago, Champaign-Urbana and a small town in Illinois. They did not gather information in suburbs.

Sandvig’s work on wireless and public policy will remain important in years to come. The recent stimulus package included $7.2 billion to expand broadband internet access, particularly in rural and other underserved areas. Wireless systems are one way that these projects will provide broadband access to the Internet. Many U.S. cities, such as Philadelphia and Long Beach, Calif., now offer free wireless network access and are considering expanding wireless reach in the future.

“In the past it’s usually been lawyers and politicians who have made decisions about wireless policy,” Sandvig says. “What we hope to do is combine engineering and social science to help policymakers make informed decisions.”


Share this story

This story was published February 13, 2013.