What cities, data and Yahoo have in common: Interaction matters

Yahoo (s yhoo) CEO Marissa Mayer caught a lot of flack for her mandate prohibiting working from home, but new research from MIT seems to back up her decision. The study, published in Nature, concludes that cities with higher rates of face-to-face interactions among citizens — what the researchers call “social-tie density” — are more productive than those where citizens interact less frequently. If valid, though, their findings could do a lot more than validate the decision by one of tech’s most-scrutinized CEOs — they could help save the world.

What the MIT researchers found, according to a press release (the actual study is behind the Nature paywall), is that their model turns out to be an accurate predictor of cities’ productivity based on gross domestic product and patent rates. They developed their model based on whatever information was available about the American and European cities they studied, based on everything from the rate of simultaneous Foursquare checkins to the contagion rates of sexually transmitted diseases — all things that suggest people being in close physical proximity.

However, lead researcher Wei Pan noted, more people doesn’t necessarily mean more social-tie density. He pointed to Beijing as an example of a city that’s growing like crazy but that’s nearly impossible to traverse from one side to the other because it keeps growing outward and there aren’t adequate options for transportation. As a result, people pretty much see the people they live near and hardly anybody else in any meaningful manner.

A sprawling Beijing from Fragrant Hill. Source: Derrick Harris

A sprawling Beijing from Fragrant Hill. Source: Derrick Harris

In stark contrast is Zurich, a city where many people outside the urban center but where a massive investment in public transportation means about 60 percent of them make their way inward during the work day. It has experienced an increase in social-tie density and productivity and even a reduction in crime.

The obligatory caveat here is that the correlations these researchers found don’t necessarily prove their theory. Other factors not analyzed could certainly be at play.

But their theory certainly seems feasible when one considers other systems where entities are forced to interact. As much as I enjoy working from home, for example, it’s easy to get caught up interacting only with my editorial colleagues (confession: we have a lot of employees on the business side, and I don’t know many of their names). I don’t think I’m alone, and I do think a lot of potentially good ideas never get a chance to formulate because our paths and our thoughts rarely cross.

There’s no doubt that companies in Silicon Valley feel this way. Massive corporate campuses are designed to maximize interactions, and even those lowly startups and private companies tend to favor open floor plans that don’t hide people behind walls or cubicles. Mayer’s mandate might suck for employees accustomed to their home-office lifestyle but — from a productivity and creativity standpoint — it never seemed too crazy to me. Earlier research from MIT suggests she’s right, too.

Inside the Googleplex. Source: Google

Inside the Googleplex. Source: Google

Or look at the concept of big data, which is as much about breaking down siloes as it is about anything else. No matter how it’s accomplished from an infrastructure perspective (e.g., throwing everything in Hadoop), taking different types of data collected for different purposes on different systems and then analyzing them against each other is one of the ways companies hope to get more from their data than just the sum of its parts. Data that doesn’t interact with other data just does its job and nothing else.

In fact, one of the major themes at our Structure conference June 19 and 20 will be the importance of building infrastructure designed to handle the ever-expanding boundaries of IT operations. How do you keep a central view of data and activity spread across numerous services and endpoints, and then how do you ultimately make sure the data and systems that need to talk to each other are able to do so?

And if you think that urban planning, corporate campuses and big data don’t have anything to do with one another, think again. The MIT research is actually a follow up to a somewhat well-known 2010 study from the Santa Fe Institute that shows how cities experience super-linear growth in everything — productiviy and patents, but also crime, disease and pollution — as they grow in population. In part, the MIT researchers wanted to figure out what was causing the increases in productivity that accompany the growth.

If there are ways to design cities that can increase productivity while keeping crime and disease in check, that would be a huge deal. If the companies and institutions that exist within these cities and that we rely on for innovation can design themselves and their infrastructure in ways that encourage productivity — and, actually, creativity (something the Santa Fe Institute’s Geoffrey West has suggested is only possible up to a point) — perhaps we can actually help solve things like crime, disease and pollution before they spiral out of control.

Feature image courtesy of Shutterstock user pio3.