Hey there, human — the robots need you! Vote for IEEE’s Robots Guide in the Webby Awards.

Close bar

Human Travel Patterns Surprisingly Predictable

Cellphone signals track travel routines

3 min read

5 June 2008—Human beings are creatures of habit, and new data shows that we may be less spontaneous than previously thought. In a study detailed this week in the journal Nature , researchers from Northeastern University, in Boston, used cellphone signals to demonstrate that human travel patterns are similar among individuals and conform to a simple mathematical model.

”We were surprised by some of the aspects of the study,” says lead author Marta González. ”There is a lot of similarity between the behavior of people.”

The study analyzed human motion by monitoring cellphone records and tracking each phone’s signal as it moved from one phone tower to the next. They used the data to come up with a probability equation for describing human movement. The results could have implications for urban planning, traffic monitoring, and the spread of disease, all of which rely on human travel.

This is not the first study to try to quantify human travel. In 2006, researchers from the Max-Planck-Institute for Dynamics and Self-Organization, in Göttingen, Germany, tracked the movement of dollar bills. As with this recent study, their data also fit into a mathematical model. But the German model could make predictions only about groups of people; it could not say anything about individual human travel. That’s because dollar bills frequently change hands. Cellphones, on the other hand, usually stay with one person, so the researchers were able to track the movements of individuals.

”In comparison to our dollar bill study, it’s a step ahead,” says Dirk Brockmann, lead author of the Max Planck study and also an associate professor of engineering sciences and applied mathematics at Northwestern University, in Evanston, Ill.

The researchers from Northeastern University obtained cellphone records from 100 000 anonymous users and tracked the data over six months. The time, date, and location of the base station receiving each call or text message was tracked, allowing the researchers to reconstruct the route of the user. Then they calculated the distances between each call.

The results showed that during the six months of the study most people traveled very short distances most of the time, while some traveled great distances. Ninety-four percent of the people traveled less than 100 kilometers, while only 0.2 percent traveled more than 500 kilometers. Short trips, naturally, were more frequent than long ones. The study also found that all users, despite the number of trips or distance traveled, visited a couple of places frequently—probably their home and their workplace.

What was more surprising was that each individual’s data fit into the same mathematical model—a type of power law—that predicts the probability of finding a person in a certain location. That probability distribution is dependent on an individual’s average travel distance and decreases the further he or she roams. A power law distribution allows for real probabilities for very large values. (A distribution of human height, by contrast, is not a power law distribution because human heights fall within a relatively limited range.)

”Human mobility and how we travel is so amazingly complex,” says Max Planck’s Brockmann. ”What is very strange is that despite this complexity, all the traveling behavior can be accounted for by very simple mathematical laws.”

Another surprising result was how little the patterns varied from person to person. ”Despite the fact that some people travel a lot and some people travel very little, when you discount the distances that people travel, [the patterns] are statistically similar,” says Albert-László Barabási, who led the Northeastern research team.

A potential application of this data is in traffic forecasting and urban planning. However, the data alone cannot predict traffic, Northeastern’s González says, and it would need to be combined with geographic information and city-specific coordinates of the routes people traveled.

Steve Gordon, a researcher at the Oak Ridge National Laboratory’s Center for Transportation Analysis, in Knoxville, Tenn., says that having a model for human movement could have profound implications for traffic forecasting.

”Currently, the only way to track cars and get an estimate of where vehicles are and of traffic conditions is with sensors,” Gordon says. And since sensors are expensive and usually only used on freeways, there is no way to monitor traffic on the major streets, he adds.

Gordon also sees potential for the model in urban planning. Knowing where people are going could help when deciding which roads need to be widened or where a new road should be built.

The next step is doing a similar experiment using GPS systems, says González. Unlike cellphones, which only show the location of the tower, a GPS system would tell the exact location of the individual.

This article is for IEEE members only. Join IEEE to access our full archive.

Join the world’s largest professional organization devoted to engineering and applied sciences and get access to all of Spectrum’s articles, podcasts, and special reports. Learn more →

If you're already an IEEE member, please sign in to continue reading.

Membership includes:

  • Get unlimited access to IEEE Spectrum content
  • Follow your favorite topics to create a personalized feed of IEEE Spectrum content
  • Save Spectrum articles to read later
  • Network with other technology professionals
  • Establish a professional profile
  • Create a group to share and collaborate on projects
  • Discover IEEE events and activities
  • Join and participate in discussions