Big Data and mobile phone geolocation data to plot the map of urban social segregation

“Cities are the most perfect social machines in existence,” claims Esteban Moro at the start of “Cities at High Resolution: Using Big Data to Understand Social Segregation,” the lecture he delivered in the BBVA Foundation as part of the series Demography Today. In his talk, he advanced some of the results of his research project at the MIT Media Lab – the laboratory of ideas run by Massachusetts Institute of Technology, where he is currently a visiting professor – in which the team is using Big Data techniques to study the intricacies of social segregation.

“Social segregation,” Professor Moro explains, “is one of the most significant processes that take place in our cities.” It comes about because people tend to live and mix with people of the same economic or social status, or who share their opinions and ideas. The problem is that traditional approaches to this demographic phenomenon based their assumptions on whether or not a place was segregated on such static variables as home address – gleaned from the census – or workplace. “Studying how society is distributed by reference to where people are at three in the morning makes no kind of sense, because most of our activities take place kilometers away from where we reside.”

New technologies are at the heart of the project. Firstly, because it draws on a massive database – over 80 billion location events of 26 million people in U.S. cities with a precision of 20 meters. “We have huge amounts of data, only accessible with today’s technology. Whole terabytes that require Big Data techniques, first to review them, then to mine them for useful information. Reaching the same conclusions without Big Data would have taken us around ten years.”

Secondly, because new mobile phone geolocation technologies permit a far more fine-grained approach. As Professor Moro puts it: “For the first time, we are able to approach the problem of segregation on a scale as tiny as the differences between one business and the business next door. Because segregation happens at that level: you have a bar on one side of the street and another just opposite, and the public varies according to the style of the place, the food on offer…” This is the first conclusion the project has arrived at, with results due out in the next few months: segregation takes place on a scale of just a few meters.

The next conclusion is that the nature of the premises (restaurants, stores, public service facilities…) occupying an urban space (a street, a neighborhood…) determine its degree of segregation. The study found, for instance, that the spaces around Chinese or ‘tapas’ restaurants display a strong social mix, i.e., a low degree of segregation. Conversely, Caribbean or Latin American restaurants attract a more segregated public.

The story is similar in the education sector: non-university centers, primary schools and high schools, are the most segregated places in the United States. “In the whole first stretch of our formal education, we only come across people of the same socioeconomic status.” This is a problem, in Moro’s view, “because all the while we are at school, we are exposed to a very limited section of society.” This intense segregation, however, eases off in later educational stages.

For Esteban Moro, the goal is plain: “We need to change these patterns of high segregation. We have to burst the bubbles. Either we do it because we realize we’re inside one, or because other people guide us out. If we always see the same faces, with the same social and economic background as our own, we become isolated, and that is a problem. What we hope to be able to shed light on is how and to what extent it is possible to break free.”

Third places

To explain how behavioral patterns are changing, the team has focused on what American sociologist Ray Oldenburg, working in the 1980s, referred to as “third places.” Esteban Moro explains it thus: “Many theories say that democracy, innovation, communities… are created in third places that are neither your home nor your workplace. It is in these kinds of places that you meet after work to exchange or invent ideas.”

In the America of the 1970s and 1980s, the typical third places were bars, hairdressing salons or bowling alleys. This model became progressively more consumption oriented over the 1990s and the first years of the new millennium, with shopping malls taking over from small local stores as the most favored third places. This was followed by another movement which saw even shopping centers relegated in importance: “People prefer to stay at home. There is a class of citizens, predominantly millennials and the preceding generation, who consume, purchase and exchange products via the digital space. Nowadays, we are not even occupying the third places that remain to us; instead they are disappearing because we would rather stay indoors and not mix with other people.”

It seems clear from the team’s research that social networks are not configuring themselves as a third place: “Our finding – says Moro – is that social networks are more segregated than society itself. In theory you can speak or interact with anyone, but in fact people are more segregated on social media than we can infer they are from their movements in the physical city. After the experience of recent years with bots and filters influencing the information we receive, it is clear that social networks have gone back to being a first place. We only hear what we agree with, and we only talk to people who think the same way as we do.”

Despite indications to the contrary, Moro is optimistic: “In the last instance, people will always need to be with each other. We still don’t know what the next third place will be, the one that replaces the shopping mall. Perhaps freeing themselves of bricks-and-mortar shopping will mean people can spend more time on open-air activities, like sports.” The full results of the project should be available before July this year.