mediaczar

Map of Porter Novelli people on Twitter on 20th Jan 2008

Three days after my last map, and after lots of internal nudging from our CMO Marian Salzman, her two helpers Tikva Morowati and Zeenat Duberia and local activists like Juriaan Vergouw, Burçu Kaptan, and Umut Ersoy, the map of Porter Novelli people on Twitter looks very different. (You can click on any of the maps in this post to go to their Flickr page where you can choose to see them at larger sizes.)

This is a map of the 156 Twitter people that I knew about this morning (there are still others coming in). The size of the blob is an indicator of how many of that person’s peers follow them on Twitter, and the colour (running from white at the “hottest” through orange and red to blue at the “coldest”) indicates how structurally important that person is. For more on this, see yesterday’s post showing the Twitter map on 17th Jan 2008.

More cohesion

Here’s what happened when I remove the ten main connectors from the old map. It falls apart.

17th Jan map (67 nodes)	10 highest betweenness nodes removed

And here’s what happens now.

20th Jan map (155 nodes)	10 highest betweenness nodes removed

You’ll see that much less damage is done now; the network is much more decentralized, thanks to a concerted effort by Marian et al. to promote cross-linking behaviours. Of course, we’re now removing around 6% of the network compared to 14%, but if the network were still as unlinked as it used to be, that would make no difference. This is on the whole, a more cohesive network.

I say “on the whole” because of the isolates (unlinked nodes) that are displayed at the top left. These are people who are new to twitter, and who haven’t yet made any friends or attracted any followers. This is one of the problems we’re trying to address; I’m considering building a “follower bot” that will help new Porter Novelli joiners (and anyone else who cares to do so) automatically follow all the other Porter Novelli twitterers (should I get over my distaste for the word “tweeple”?)

Weaknesses

As Tikva and others have pointed out, simple linking isn’t enough to really rank people in this network. You’d want to look at other quantitative metrics. Recency, frequency and tenure spring to mind, of course; but I’d also want to see things like who (in the network) is most often addressed or quoted using the “@<username>” convention. We’ve been building an “eavesdropping” bot to help us get a grasp of that sort of thing, but we’re still not there (the code works, but I can’t get it to run on my server yet…)

On the qualitative side, we’re also missing important things like “how much one person likes, is indifferent, or dislikes another” (there has to be a better way of expressing that.) I don’t really do qual measurements at present; I want to be able to get a whole load of stuff out of the way before I even begin to deal with tricky stuff like that. Is this an Aspergers thing, do you think, or am I just being male?