Great stuff; next thing on my todo list was to look at the *all* the people the journalists follow, to see if they cluster too, or whether the journalists follow a diverse set of people. (I’m not sure how big this network gets?). The other thing is to look at the distribution of followers, but I suspect that network will be hundreds of thousands of nodes, which my machine may choke on! I’m not sure how well Gephi scales once you get past tens of thousands of nodes?)

Something else on my to do list to is grab the links (and maybe hashtags?) shared by folk on the lists and then treat these as “documents” so that they can be visualised by Gource using its Custom Log Format (e.g. http://pyevolve.sourceforge.net/wordpress/?p=1423 ) I’m not sure if this will work, but it’s an interesting experiment for a spare few hours…;-) Gource typically visualises users who check code files into a repository. So I wonder, if we treat sharing a link as if it were a file checkin, we should be able to visualise who’s sharing the same links? If we can also colour the individual journalists by media group, we get to see the extent to which they share different links. (When visualising the links, we need to account for different shortenings of the same link, so we probably need to resolve them?)

Something else I’ve explored is to look at who’s tweeting the links a journalist has tweeted (if they’re shortening links as a logged in bit.ly user). So for example, this image: http://www.flickr.com/photos/psychemedia/5437539168/in/photostream shows connections between people who tweeted one or more of 15 links that also happened to be recently shortened on bitly by @charlesarthur. This used the bit.ly api to pull the most recent links @charlesarthur had shortened, and then the Backtype API to see who had tweeted each of those links. (See http://blog.ouseful.info/2011/02/03/visualising-ad-hoc-tweeted-link-communities-via-backtype/ for a quick discussion about Backtype.)

PS I’ve yet to tap into the firehose, so will pick up on your howtos:-)

I’ve stumbled upon your visualisation on guardian.uk. I took the liberty to create something similar for the american journalists on twitter. I’ve took those that can be found on http://muckrack.com/sources.

Here is a nice visualisation of their friends follower connections. Node size is by directed page rank and color by organisation:

http://zoom.it/7601

You can find more info on the set on my page dedicated to research on twitter:

https://sites.google.com/site/twitterresearch09/articles/journalistslikejournalists

I’ve also used technique to plot how folk using the same hashtag, or sharing the same link, are connected by friend relations, which can be an aid to discovering people (and hence resources, via the links they share) with a certain amount of “influence” in a particular subject area. If we look at folk using the hashtag and their friends who don’t, the technique can also illustrate the extent to which a conversation e.g. around a hashtag is going on in an echochamber.

I suspect the reason that it was widely tweeted was that lots of journalists (with large follower numbers) tweeted it because they were on it… ;-) Hmm… when I get a chance, I think I’ll graph the connections between folk who did tweet it…