Can the laws of physics untangle traffic jams, stock markets, and other complex systems?
Rochester professor Gourab Ghoshal is using the laws of physics to untangle the complex systems behind human behavior, urban planning, and social networks.
Today, the increasing availability of digital traces provides researchers like Ghoshal and the members of his lab an incredible amount of data for complex systems research. Using GPS location tracking, check-ins on apps like Foursquare, geocaching from Twitter posts, and, under some circumstances, call data records from cell phones, they are able to find patterns in human mobility, traffic, and disease progressions with greater accuracy and precision than ever before. But the amount of data available also means that the burgeoning field of complex systems can be overwhelmed without some constraints.
“You’re going from orders of tens and hundreds of data points to orders of millions and billions,” Ghoshal says. “When you have this change of scale, the simple ways of drawing diagrams don’t work; you need laws to dictate how the data interact.”
That’s where physics enters the picture.
Ghoshal and members of his lab used science to untangle complex systems and networks. Their work extends to three main branches: human dynamics and behaviors; urban systems and city science; and social networks. Applying the universal rules of physics and mathematics, they are able to build foundational models onto which other researchers—such as economists or city planners—can introduce more complex variables.
There are differences between the way a physicist looks at a problem compared to a computer scientist,” says Hugo Barbosa, a post-doctoral researcher in Ghoshal’s lab, whose PhD is in computer science. “Physicists are more interested in the fundamental rules, the things that are universal, regardless of the populations. They want to understand the basic components of those models and make those components as general and universal as possible.”
One project, conducted by Surendra Hazarie, a PhD student in physics, uses the data to look at segregation patterns in cities, along lines of income, race, or other characteristics. “By looking at the ways populations flow around cities in the different regions, we can examine the way those populations are split up in important ways,” he says. “Maybe wealthier communities tend to box themselves in and things like that.”
It’s more about how can we leverage this knowledge to, for example, improve the synchronization of traffic lights and diverge the flow of people to a different area of the city if there is an accident? How can we foster and promote reducing segregation? How can we help prevent the spread of diseases in existing airports or hospitals?”
Understanding how people and locations are connected and analyzing those relationships through network centrality helps inform my predictive models in ways that would not have been possible using traditional statistical methods,” Marshall says. “Network science changes the way we view patient movements in the hospital and helps us develop tools to improve patient outcomes and reduce the spread of infection.”