Skip main navigation
We use cookies to give you a better experience, if that’s ok you can close this message and carry on browsing. For more info read our cookies policy.
We use cookies to give you a better experience. Carry on browsing if you're happy with this, or read our cookies policy for more information.

Numerical analysis of a community

Perhaps the most obvious way to benefit from social media analysis is to use it to study the way that individuals and communities behave. It should also make us think more carefully about our own contributions to social media, as often our individual posts are forming a bigger picture about our views and behaviour.

Social scientists have been studying these interactions for many years, but because the social web is machine readable we can apply new methods and this gives us new opportunities to undertake quantitative analysis of these networks and to build numerical and statistical models.

For example, scientists from Microsoft Research have looked at whether user communities in an online forum (they were looking at Usenet, a pre-web bulletin board system) took different shapes depending on the subjects being discussed (Fisher et. al 2006). In this case two individuals are connected if they have interacted on the forum (for example, responding to a post).

They found that different types of forum contained different types of behaviour, for example technical forums had many question and answer style interactions, with a group of experts providing many of the answers (individuals of high degree frequently responding to individuals of low degree). On the other hand support forums contained more discursive interactions and were more uniform (most individuals had a similar degree).

Adamic et al. (2008) undertook a similar analysis of posts on Yahoo Answers, using a number of numerical statistics to describe what was going on in different areas (for example, average thread and post length). They defined entropy as a number that captures how focused on a topic a given user is (a user who only posts in one category has entropy of 0), and showed that there was a relationship between low entropy and a high quality of answers in technical categories.

These kinds of studies show that network analysis can give us important numerical evidence about how people behave.

What sort of numerical analysis might you apply to the social networks that you are involved with, and what are the interesting things that you might look for?

Further reading

Optionally you might like to read the summary of the Adamic paper on Yahoo Answers that is attached to this article.

It is a good example of how you can build interesting new properties (in this case entropy) from existing well-established ones. It also shows how you can use low-level statistical and analytical methods to answer high-level questions.

References

  1. Fisher, D. (2006). You Are Who You Talk To: Detecting Roles in Usenet Newsgroups. Proceedings of the 39th Hawaii International Conference on System Sciences

  2. Adamic, L. A., Zhang, J., Bakshy, E., & Ackerman, M. S. (2008). Knowledge sharing and yahoo answers: everyone knows something. Presented at the WWW ‘08: Proceeding of the 17th international conference on World Wide Web, ACM.

Share this article:

This article is from the free online course:

The Power of Social Media

University of Southampton

Get a taste of this course

Find out what this course is like by previewing some of the course steps before you join:

  • Welcome to the course
    Welcome to the course
    video

    This course explains how by understanding networks better we can exploit social media. Watch Dr David Millard & Dr Lisa Harris tell you more.

  • What are networks?
    What are networks?
    video

    Watch Dr David Millard explore what we mean by a network, and look at some of the examples of networks around us and their characteristics.

  • Recruitment in a digital world
    Recruitment in a digital world
    video

    Watch Dr Lisa Harris, Nic Fair and Sarah Hewitt discuss the role of social media in recruitment from the perspective of employers & potential employee

Contact FutureLearn for Support