Tweety pi?

Those of you who subscribe to the University’s Twitter feed may have spotted a tweet last week that read “Maths researchers @UniStrathclyde work with industrial partner on new way to measure impact of social media campaigns”. If you chased this up, you might have found your way to a Youtube video in which our own Professor Des Higham sings the praises of the collaboration but doesn’t get much chance to talk in detail about the maths. Fortunately, Des was happy to point DoF in the direction of some sources that give a bit more information for anyone who’s interested…

A good place to start might be a recent conference paper called Twitter’s Big Hitters. This explains the basic ideas: with social media like Twitter becoming an increasingly important part of the way in which companies market themselves and their products, it’s becoming increasingly important for the marketers to track and influence what’s going on in them. Put crudely, if you can find out who everybody else is following and convince these cool kids of your worth, they will then do a lot of your marketing for you.

We can represent a social network mathematically as a type of graph: a set of nodes connected to each other by directed edges. For example, the nodes might be Twitter users, and a directed edge might be added, for example, every time one user follows another user. There are some reasonably well-developed metrics for describing how well-connected two nodes in a network are, and thus for determining which nodes are the most central or the most important. (The celebrated Bacon number and Erdős number are early and relatively simple examples of this game, played with movie actors and mathematicians respectively.) What Des and his collaborators at the University of Reading and in the Bloom Agency in Leeds have done is to extend these metrics to networks that evolve in time, rather than taking a static “snapshot” of the state of the network. In a rapidly changing system like Twitter, it’s plausible that this could make a substantial difference to which nodes — which Twitter users — appear to be most influential, and that’s exactly what Des and his collaborators have found. If that description whets your appetite, there’s a more substantial account available in a Department Research Report, Dynamic Targeting in an Online Social Medium.

The work on Twitter is in turn related to a wider collaborative project called MOLTEN (no, I don’t know whether the project or the acronym came first), which applies similar ideas in a wide range of contexts. If you find the idea of ranking people by their social networking activity disturbing or even objectionable, you might be more comfortable knowing that related techniques can be applied to analyse processes observed in the human brain, or to find new ways to defeat viruses that attack mobile phones.

That’s the nice thing about applied maths — everything, it sometimes seems, really is connected to everything else…


This entry was posted in Articles. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s