With the popularity of Facebook, Twitter, LinkedIn, and other social networks, we're increasingly defined by who we know and who's in our network. These websites help us manage who we know—whether personally, professionally, or in some other way—and our interactions with those groups and individuals. In exchange, we tell these sites who we are in the network.
These companies, and many others, spend a lot of time on and pay attention to our social networks. What do they say about us, and how can we sell things to these groups?
In this chapter, we'll walk through learning about and analyzing social networks:
- Analyzing social networks
- Getting the data
- Understanding graphs
- Implementing the graphs
- Measuring social network graphs
- Visualizing social network graphs
Although the Internet and popular games such as Six Degrees of Kevin Bacon have popularized the concept, social network analysis has been around for a long time. It has deep roots in sociology. Although the sociologist John A. Barnes may have been the first person to use the term in 1954 in the article Class and communities in a Norwegian island parish (http://garfield.library.upenn.edu/classics1987/A1987H444300001.pdf), he was building on a tradition from the 1930s, and before that, he was looking at social groups and interactions relationally. Researchers contended that the phenomenon arose from social interactions and not individuals.
Slightly more recently, starting in the 1960s, Stanley Milgram has been working on a small world experiment. He would mail a letter to a volunteer somewhere in the mid-western United States and ask him or her to get it to a target individual in Boston. If the volunteer knew the target on a first-name basis, he or she could mail it to him. Otherwise, they would need to pass it to someone they knew who might know the target. At each step, the participants were to mail a postcard to Milgram so that he could track the progress of the letter.
This experiment (and other experiments based on it) has been criticized. For one thing, the participants may decide to just throw the letter away and miss huge swathes of the network. However, the results are evocative. Milgram found that the few letters that made it to the target, did so with an average of six steps. Similar results have been born out by later, similar experiments.
Milgram himself did not use the popular phrase six degrees of separation. This was probably taken from John Guare's play and film Six Degrees of Separation (1990 and 1993). He said he got the concept from Guglielmo Marconi, who discussed it in his 1909 Nobel Prize address.
The phrase "six degrees" is synonymous with social networks in the popular imagination, and a large part of this is due to the pop culture game Six Degrees of Kevin Bacon. In this game, people would try to find a link between Kevin Bacon and some other actor by tracing the films in which they've worked together.
In this chapter, we'll take a look at this game more critically. We'll use it to explore a network of Facebook (https://www.facebook.com/) users. We'll visualize this network and look at some of its characteristics.
Specifically, we're going to look at a network that has been gathered from Facebook. We'll find data for Facebook users and their friends, and we'll use that data to construct a social network graph. We'll analyze that information to see whether the observation about the six degrees of separation applies to this network. More broadly, we'll see what we can learn about the relationships represented in the network and consider some possible directions for future research.