Intuitively, the close community spirit of village life and the crowded bustle of the big city suggest very different qualities of social life. A paper published today in the Journal of the Royal Society Interface finds, however, that the social networks of city dwellers are not so different from those of village dwellers.
Scientists from SFI and the MIT Senseable City Lab teamed up with researchers from British Telecommunications and Orange Labs to examine social relationships of people living in towns and cities in Portugal and the UK.
Sociologists have long been interested in the way cities affect social interactions, but previous methods of quantifying these interactions relied on traditional surveys – essentially asking people who they talked to and how often, a time-consuming approach that depends heavily on the accuracy of the respondents' answers.
Today, telecommunications companies keep records of every call made by every customer – information that includes call time, duration, location, and numbers dialed. These data, stripped of all identifying personal information, offer much more extensive records of social interactions than is possible through traditional surveys.
As the basis for their study, the researchers obtained a dataset comprising the majority of landline calls made in the UK during a one-month period in 2005, and another dataset of millions of mobile phone calls in Portugal during a 15-month period in 2006 and 2007. Together, these two datasets represent several billion phone calls.
“This is unprecedented raw material for analysis, poised to change our understanding of society – potentially opening the way to something that some people have started calling ‘computational social sciences’,” says Carlo Ratti, director of the MIT Senseable City Lab, one of the paper’s co-authors.
The researchers then constructed networks of phone interactions for each town in the UK and Portugal, with each individual user represented by a node, and connections between them indicated by a link.
The team’s analyses of the network data revealed that the number of calls made by an individual, as well as his or her total number of connections, depends on the town's size according to a mathematical relation: the larger the town you live in, the more people you call and the more calls you make. Interestingly, this relation is "superlinear," meaning that on average, as the size of a town doubles, the total number of social interactions will more than double in a predictable way.
The team also found, however, that the group clustering of social circles (the odds that your friends mutually know one another) does not change with city size, regardless of whether you live in the five square-mile town of Lixa in northern Portugal or the bustling capital city of Lisbon.
Social networks derived from mobile phone data in Portugal: An average urban dweller of Lisbon has approximately twice as many contacts than an average individual in the rural town of Lixa. However, the odds that an individual's contacts mutually know one another remains largely constant. (Credit: Kael Greco, MIT Senseable City Lab)
The findings point to the conclusion that human beings living in small towns and large cities alike instinctively form tight social communities. But if you live in a small community, your social circle is more or less determined by those who live around you, whereas in a large city you have more freedom to select which of the thousands of people around you will constitute your social circle.
“It seems that even in large cities we tend to build a tightly-knit community, or ‘village,’ around ourselves”, says Ratti. “In a real village, connections might be defined by proximity, while in a large city we can elect a community based on affinity, interest, or sexual preference, for example.”
“People tend to think of cities as people, buildings, roads, pipes, and so on,” he says. “But at a more fundamental level, cities are really about connections. These connections form networks of people and organizations that enable the production of all products of civilization, from modern economies and fast innovation to complex bureaucracies and political institutions.”
“That social interactions per person increase with city size begins to explain how so many socioeconomic quantities, from GDP to violent crime, scale superlinearly,” he adds. “We had developed theory that predict the superlinear growth of social connections in the way we observe here, but this is the first time that we can observe this phenomenon directly and explore it in detail! It is tremendously exciting.”
Markus Schläpfer of MIT’s Senseable City Lab and SFI, the corresponding author on the paper, says the team’s findings have important implications for the way information and ideas diffuse throughout a city. Ultimately, this may also help researchers understand phenomena such as the prevalence of certain contagious diseases.
“This was an incredible opportunity, made possible by today’s widespread use of mobile communication technologies”, he says. “Data of this type keep getting better and better. It will be extraordinary to use them in the future to see how cities around the world reproduce the patterns we observed for Portugal and the UK, and watch fast-growing cities develop as immense social networks. It throws open lots of possibilities to study the organization and dynamics of entire cities.”
Co-authors of the paper, titled “The scaling of human interactions with city size,” include Schläpfer; Bettencourt; West; Ratti; Mathias Raschke of Raschke Software Engineering, Sébastian Grauwin of the Senseable City Lab, Rob Claxton of British Telecommunications, and Zbigniew Smoreda of Orange Labs. The project was developed as part of Ericsson’s “Signature of Humanity” project.
Read the paper in the Journal of the Royal Society Interface (July 2, 2014, subscription required)
Read the article in New Statesman (July 17, 2014)
Read the article in Spiegel (July 7, 2014)