Picking up on Brian Kelly’s traffic seeking #Sherlock post (Isn’t #Sherlock Great! (TV & a ‘Second Screen’ For the Twitter Generation)), I just made a quick tweak to my emergent social positioning code to have a peek at who’s commonly followed by the folk who’ve recently been using the #sherlock tag on Twitter.
By running a search at around 10.15pm for the most recent 1500 tweets using the #sherlock hashtag I grabbed a list of tag users; I then filtered this down to select users who had used the tag twice within the sample set, leaving a set of about 250 users or so. I then found the friends of this sample set (that is, lists of all the people they follow) and constructed a graph of folk followed by a sample of hashtag users. I then filtered this graph to nodes with a degree of 75 or more, a crude way of capturing folk who are followed by a significant number of the tag using sample set and excluding folk in the tag sample set who only follow a few folk.
Here’s the result – the quick’n’dirty interpretation is that it gives a quick sketch of which folk on Twitter are followed en masse by folk who were using the #sherlock hashtag (subject to all sorts of sampling errors…which is why it’s just a sketch…).
If we relax the network degree filter to 25, this is what we get (nodes sized according to eigenvector centrality and coloured according to groups identified using the Gephi modularity cluster statistic):
PS it’s also possible to search for a URL on twitter to see who’s been tweeting it recently… Running a search for the most recent (23.40) 500 tweets containing a link to the blog of Dr John Watson, I discovered 320 or so twitter users. Projecting from them to their friends and filtering the resulting graph to nodes of degree 20 or more, here’s how Dr Watson’s blog is socially positioned in the sense of mapping folk who are commonly followed by folk who have recently tweeted a link to the blog.
THe question is: what’s this actually useful for?
PS see also Social Media Interest Maps of Newsnight and BBCQT Twitterers which starts to explore ways of comparing interests of folk tweeting around different hashtags.