AAAS 2013: Predicting human behavior

Photo via Thinkstock.
Photo via Thinkstock.

Photo via Thinkstock.

If you’ve driven on the highway, you’ve seen it: The traffic jam appears out of nowhere and disappears just as mysteriously.  We blame the cars around us for their poor driving skills, and slam on our own breaks. During an AAAS annual meeting session hosted by Northeastern professor Albert-László Barabási, Northwestern professor Dirk Helbing showed a video of cars driving at constant speed around a circular race track. Eventually, small variations in driving speed propagate into a cascading traffic jam that travels backwards around the track.

You can imagine how more cars would allow this phenomenon, which emerges as a result of the cars’ interconnectedness, would happen a little quicker on a more congested track. This is how I think about our world these days. More than seven billion people are traveling around this giant track and we’re getting ourselves into some nasty jams. Climate change, financial meltdowns, the Arab Spring, SARS: most of the large scale problems we’re now facing stem from the fact that our system has grown too big for its britches.

Of course, it’s not just a problem of size. It’s also a problem of connectedness. Because human socioeconomic and technical systems are deeply intertwined through global networks of things like mobility and money, small perturbations are vastly amplified.

But just as our population has grown, so too has our ability to track that population. More data has emerged in the last two years than did in all preceding human history. Alone that would just be an overwhelming concept worthy of little more than a minor panic attack, but coupled with insights from the physical sciences and mathematical modeling, we are now in a position to actually probe all that data and use it to understand the underlying mechanisms that govern human behavior.

The session was titled “Predictability: From Physical to Data Sciences,” and five speakers in addition to Helbing discussed ways in which human behaviors can be forecasted much the same way we track the path of a hurricane.

Northeastern research assistant professor Chaoming Song looked specifically at the predictability of human motion. Using cell phone data, Song and other member’s of Barabasi’s lab showed that human mobility is an interplay between two phenomena: returning and exploring. Most of the time, we’re moving from one familiar place to another. Every now and then, we venture off to some new destination. Intuitively, it’s not that surprising, but the fact that Song and his colleagues are now able to model those patterns with extreme accuracy is ground breaking.

Work like this informs the work that Northeastern professor Alessandro Vespginani and Northwestern professor Dirk Brockman are doing to track the spread of epidemic diseases around the globe. Vespignani’s GLEaM model compiles a series of layered human mobility data sets to predict epdidemic trajectories and severity. His work recently demonstrated that in order to have any kind of effect, transportation restrictions for containing diseases would have to shut down 99% of human movement in order to be remotely effective. In contrast to the way small variations in speed amplify into major traffic jams, large interventions can sometimes have minimal effect in a complex system.

Instead of using real data about human motion like Vespignani, Brockman looked at the movement of dollar bills around the globe using the website wheresgeorge.com. Brockman used his proxy data set to demonstrate where and when diseases would break out next and despite the seemingly random pattern, they results were strikingly similar to those derived from Vespigani’s method. This urged Brockman to search for a “more fundamental” approach. He decided that perhaps our global connectivity means that the distances between two cities shouldn’t be described in geographic terms, but rather “effective” terms. If the fraction of people moving between two airports is enormous, those two airports should be considered as very close together. But if no one travels between them, they are very far apart (even if geographically they’re quite close).

He redrew the network of international airports using this new definition of distance and now diseases spread just as you might expect, like a ripple propagating through water, concentric circles emanating form a central focal point like a pebble or the initial case of SARS.

Vespignani noted that in the 14th century it took several years for the black plague to ripple through the population, but it did so just like that: with a wavelike motion, geographically propagating from one city outward. Brockman’s approach shows how our interconnectedness changes only the shape of the network on which that ripple travels.

Other speakers at the session included Marta Gonzalez of MIT and Boleslaw Szymanski of Rensselaer Polytechnic Institute.