Archive

Posts Tagged ‘Stanley Milgram’

More about Social Networking tools

In my last post I wrote a little about networking tools like Facebook and LinkedIn. I also posted a picture of the network of my friends, and how they are connected between themselves. Each of the dots in the graph represent a friend, and the edges represent a connection between two of them. Stronger colored connections means that the people also have other similarities besides being friends with me and each other – like being fans of the same thing or belonging to the same group.

Here is the same graph arranged in a different way, and with light colors:

My Facebook friends

My Facebook friends

This form of the graph shows the clustering better, but the individual edges between the persons are harder to see. In the application, you can hover any one of the dots, see who it represents and also see which of the other dots are connected (that is, which friends that person has in common with me). Naturally there are no dots representing myself in the graph, since that dot would be connected to every other.

I found this interresting application through the article “The convergence of social and technological networks” by Jon Kleinberg, posted in the November 2008 issue of Communcations of the ACM. This is a very interesting article describing how small the world has become due to these social networking tools (at least for those using it). Just think about how fast some things spread on sites like this – be it a funny video or a rumour, or some strange applications.

When you look at the picture above, you see more than one cluster. When one of my friends on Facebook post something, it will appear in my news stream. If I then like it, I may comment on it, or maybe even post it myself. Or if it is a group someone joins, I join it to. Many of the others in the same cluster as the initial poster have seen it before, and so they will then see it again – but the interesting thing is that so will all those of my friends who are in different clusters (and of course those in the same clusters with no direct link to the initial poster). So when my Facebook friends also sees this – they might find it interesting too (given that any of my friends share my interest – which isn’t that unlikely), and so the snowball can roll on.

The article mentions how this reminds us of the concept of “six degrees of separation” – a term originating in a play by John Guare, based on the 1960’s research of Stanley Milgram. In this research, people were asked to try to forward a message to a given person which they did not know, only by passing it through people they knew. It turned out that the median of steps the message went was only six.

I’ve often heard this theory reffered to – people saying that “you can get to anyone in the world by just six steps!” – it is probably an exaggeration, at least if you base it on the research (though I think it’s thought to be an average – implying of course that there are many people who are more than six steps away since there areĀ  a lot of people less than six steps away) – but often you will see that the number of steps to go from one individual to another is surprisingly few.

Kleinberg’s article tries to explain that by using the fact that even if most of our friends are close to us (in some way – be it geographical or based on interests) – there are always someone who has some odd connection. Look for instance on the connection I’m making between the people in my church (the largest cluster) and the people I went to folk high school with (the densest cluster). In my Nexus graph there are no edges between these clusters – indicating that I’m the only person who has Facebook friends in both groups (as far as I know I’m the only one who knows people in both these groups – with one possible exception, but they’re not connected on Facebook).

In the news article “Analyzing Online Social Networks” from the same issue of Communications, Bill Howard speculates that the average step to get from one random individual to another in the online networks is more likely less than three.

I recently joined the professional network tool LinkedIn. This tool provides some statistics about this. I’ve now got 25 contacts in LinkedIn – some of these are about as new as me, while other has been on there for a long time, and have a lots of contacts. According to the statistics, there are 1000+ people two degrees away (contacts of my contacts), and 126,400+ people three degrees away. So out of the 35 million LinkedIn users, I am able to get to 127,500+ people two steps or less.