A couple of weeks ago, I put started cobbling together some clunky scripts to collate network data files from lists of people twittering with a particular hashtag (First Glimpses of the OUConf10 Hashtag Community). I’ve got a Twapperkeeper key now, so the next step is to pull archived hashtagged tweets from there to generate my hashtaggers list, and then use that data as the basis for pulling in friends and followers links for particular individuals from the Twitter API.
One thing I’d like to start pulling together is a set of tools for providing network and backchannel analysis around hashtag communities. Andy Powell has already published a site that summarises hashtag activity in the form of Summarizr using a Twapperkeeper archive:
So what else might we look for?
Mulling over my own Personal Twitter Networks in Hashtag Communities, the metrics I report include:
– Number of hashtaggers [Ngalaxy]
– Hashtaggers as followers (‘hashtag followers’) [Gfollowers]
– Hashtaggers as friends (‘hashtag friends’) [Gfriends]
– Hashtagger followers not friended (‘serfs’) [Gserfs]
– Hashtagger friends not following (‘slebs’) [Gslebs]
– Hashtaggers not friends or followers (‘the hashtag void’) [Gvoid]
– Reach into hashtag community [Greach=Gfollowers/Ngalaxy]
– Reception of hashtag community the proportion of the the hashtag community that are followed by (i.e. are friends of) the named individual; [Greception=Gfriends/Ngalaxy]
– Hashtag void (normalised) [Normvoid=Gvoid/Ngalaxy]
– Total personal followers the total number of followers of the named individual [Nfollowers]
– Total personal friends: the total number of friends of the named individual [Nfriends]
– Hashtag community dominance of personal reach: the extent to which the hashtag community dominates the set of people who follow the named individual, [Domreach=Gfollowers/Nfollowers]
– Hashtag community dominance of personal reception: the extent to which the set of the named individual’s friends is dominated by members of the hashtag community, [Domreception=Gfriends/Nfriends]
Anyway, it strikes me that calculating those measures as means (and standard deviations) across all the members of the network, along with more traditional social network analysis network centrality or clustering measures, might help identify different signatures relating to the maturity of different hashtag communities (for example, the extent to which they are just forming, or the extent to which they have largely saturated in terms of members knowing each other).
These metrics might also change over the course of an event being discussed via a particular hashtag.
OUseful.Info, the blog... 
Can you pick up who is adding/added to friends/followers at what point?
If you look at TwapperKeeper, you’re taking a retrospective view – which may show the network as more saturated with connections than it was before the hashtag event. Is it possible to snapshot friends/followers when a person first uses that hashtag, and then to snapshot them again at the end of the event?
This could help indicate which hashtags extend communities and which strengthen existing links.
This approach was taken during Dev8D earlier this year to produce the Dev8D community network growth maps ( http://blogs.ecs.soton.ac.uk/webteam/2010/03/03/a-first-look-at-the-dev8d-twitter-network/ and http://blogs.ecs.soton.ac.uk/webteam/2010/03/04/dev8d-twitter-network-part-2/ ). I keep meaning to chat to @andypowe11 to see if we might be able to find a way of grabbing this sort of data in a Summarizr context over a specified period of time (such as from the week before an event to a couple of weeks after it).