Structure of network drives friends to congregate into many small, highly interconnected communities

Nov 08, 2012

For the first time, the dynamics of how Facebook user communities are formed have been identified, revealing surprisingly few large communities and innumerable highly connected small-size communities. These findings are about to be published in EPJ Data Science by Italian scientist Emilio Ferrara, affiliated with both Indiana University in Bloomington, Indiana, USA and his home University of Messina. This work could ultimately help identify the most efficient way to spread information, such as advertising, or ideas over large networks.

No previous work has attempted to analyse the of as a proxy to understanding real world communities at the same scale.

The author elected to analyse Facebook with the typically used to study complex systems in order to uncover its dynamics. First, Ferrara acquired a snapshot of the structure of the users' using several techniques of statistical sampling applied to the anonymised public profiles of Facebook users. He then validated his approach to detect communities by comparing the outcome of several statistical methods and by using various algorithms.

He found that Facebook communities emerge as a result of the network's structure, which is based on creating networks of friends. It therefore has little to do with how individual users behave. Ferrara also realised that only few large communities emerge. Instead, users tend to aggregate in small-sized communities that are extremely interconnected. This type of structure is known to optimise the efficiency of communications among users. Indeed, short paths of communication can connect any pair of users, even if they belong to completely disparate communities.

Ultimately, this approach could be applied to verify a social theory known as Granovetter's "strength of weak ties", whereby loose interconnections among users yield better opportunities and more efficient communication channels.

Explore further: Communication-optimal algorithms for contracting distributed tensors

More information: E. Ferrara (2012), A large-scale community structure analysis in Facebook, EPJ Data Science 1:9, DOI 10.1140/epjds9

add to favorites email to friend print save as pdf

Related Stories

The modeling of multiple relationships in social networks

Nov 17, 2011

A study conducted by Columbia Business School's Prof. Asim Ansari, William T. Dillard Professor of Marketing, Marketing, and Oded Koenigsberg, Barbara and Meyer Feldberg Associate Professor of Business, Marketing, alongside ...

Facebook tops 350 million users, tightens privacy

Dec 02, 2009

Facebook is enhancing privacy controls and eliminating its regional framework for online communities as the Internet's most popular social networking service tops 350 million users.

Harnessing the predictive power of virtual communities

Jan 30, 2012

Scientists have created a new algorithm to detect virtual communities, designed to match the needs of real-life social, biological or information networks detection better than with current attempts. The results of this study ...

Recommended for you

Designing exascale computers

Jul 23, 2014

"Imagine a heart surgeon operating to repair a blocked coronary artery. Someday soon, the surgeon might run a detailed computer simulation of blood flowing through the patient's arteries, showing how millions ...

User comments : 1

Adjust slider to filter visible comments by rank

Display comments: newest first

mildly vexed
not rated yet Nov 12, 2012
This was done a few years ago at another social network, with similar results. Nice to see someone getting to publish on the subject, because it's becoming increasingly important, not only in designing to empower these behaviors, but in letting people manage their weak ties well and safely, and not get unfortunately isolated by their strong ties. (Seems to me some of that insight could help understand recent political events in the U.S., too.)

Please ignore my user name - I started off mildly vexed someone else got to publish, but I've ended up happy the info is out there :-)