An acquaintance on facebook tipped me off to NodeXL, a tool for analyzing networks. I loosely knew of this, having seen it pop up in a methods section here and there. But I had no idea that it made it so easy to gather realtime updates from Twitter streams. This is mostly for my future memory.

Here’s an example of using NodeXL to develop a graph of the community of users following @WiredUK, complete with step by step instructions. Here is the resulting chart:


Other interesting, related links:

  • Presentation using NodeXL
  • NodeXL is a plugin for Excel for Windows only.  Us Mac users can either run Parallels, which can bog things down, or use this innovative method of running it on a virtual server on Amazon EC2.
  • Calculating geometric mean of two users
  • Word association networks
  • NodeXL Graph Gallery
  • A whole bunch of posts about it
  • Geocoding tweets
  • And a book on using it

Of course, Gephi is the prettier version of this, and some recommend exporting data to that tool.

Twitter’s Future Direction

What’s in a Tweet?

As it’s been filled up with gooey media, Twitter’s stream has felt more and more like Facebook’s, but it’s inverting Facebook’s model here by allowing its media to be disembodied and spread anywhere — you don’t have to go to Twitter to look at a post, unlike Facebook. Twitter isn’t a closed stream in this model so much as it is a kind of ubiquitous media platform that can hold anything and be anywhere, all connected to the backbone of the stream.

The direction Twitter is taking is interesting: more hub like and open (except to developers) than facebook, with the clear goal of being the rivet through which your stream flows. And it’s not closed off, like Facebook.

PotD – 01/25/10

Shared Experiences

I’ve been reading this evening about the development and study of shared memories (Halbwachs, Pennebaker and Weeden). I am sincerely intrigued by the notion of shared memories: that there are constraints imposed by society that shape how an individual person remembers things, and that as people tend to interact with their peers, those social norms are both reinforced and altered in subtle ways – thereby impacting the individual’s memories, since memories are always viewed through the lens of the present. Additionally, major, shared events like 9/11 or the fall of the Berlin Wall have different impacts on different generations: most notably, if the event occurred while you were between the ages of 12 and 25, you’re more likely to remember many details of it, in fact, to have a flashbulb memory – one in which you remember where you were when you heard about it. Obviously, for most Americans, the big recent events were 9/11 and Obama’s inauguration.

But I’m really interested in sub-cultures and how they develop, especially political sub-cultures. Often, these can be ethnically or regionally based, like the varying ethnic groups in the former Yugoslavia or, say, the South today. Or they can be political sub-cultures, like the online left leaning communities surrounding websites like Daily Kos or those fetishizing Glenn Beck’s musings. (Was that a little biased? See, that’s why I’m not studying AP).

As these sub-cultures evolve, they develop their own cultures. According to Weeden, whose interpretation I favor, culture is just a set of practices and signs and symbols that are intelligible to the relevant group, and are thus subject to fluid reinterpretation and change. This is great when the sub-cultures are internet-based fandoms or idolation of Emile Lagasse. But when they focus on the political realm, they tend to impact how someone views the world. As the community deliberates, each individual’s memory of an event, be it the stimulus package or the attempted Christmas Day bombing, is affected by that conversation. New norms develop that are not necessarily based on what actually happened, but on the reinterpretations of those events by the community (there’s a whole other question about whether those interpretations primarily occur at the elite level and trickle down or at the grassroots level and percolate up, but I’m going to sidestep that one for now).

So what happens when political discourse is relegated to two teams that mostly only talk to themselves? Well, hmm – “death panels”? The unquestioning belief that the stimulus was more damaging than helpful? The “Cornhusker Kickback” – a term I’m sure 75% of the country, including a good chunk of those that are politically involved, have never heard. I’m sure there are a number of these on the left as well, but I don’t feel like looking them up.

The point is relatively clear, and not all that novel: polarization, combined with the new marketplace for ideas, allows people to conduct a political conversation only with those who think like they do. So while the people in the picture above are all sharing the experience of seeing the Cardinals kick some old man Favre ass, there isn’t a corollary in the political arena, because interpretations will soon branch off into partisan and polarized directions (again, the same thing happens in sports – it just doesn’t matter as much in terms of tangible effects on people’s lives).

The point of all this is that I wanted to describe a research idea. People in my line of research have seen this visualization, which succinctly shows the relationship between blogs. But it does this at a single historical moment – now – and not through history. In other words, it doesn’t show the development of those links – if left-leaning blogs slowly link more and more to left-leaning blogs and right-leaning blogs to their peers, etc.

The second part of this research project involves tracing not just the links to other websites but the actual positions taken by the blog authors themselves. (This was inspired in part by this article chronicling the fight between Little Green Footballs (LGF) and its former ideological peers in the warblogger community (incidentally, some of the websites mentioned were objects of a section of my MA thesis, and yes, some of the namecallling was pretty disturbing).

So the question, then, is as follows: do online blogging communities tend to decrease their links to ideological foes as their ideologies gain rigidness? In other words, is there a correlation between online communities converging on an ideological center and a relative increase in outbound links to other sites within the community? I suppose the ideological convergence doesn’t have to mean ideological conformity, maybe just drifting toward the mean on the part of a given blogger would count.

I think they will. (Of course, LGF’s ideological shift is actually an exception to my theory.)

Now who wants to give me time/funding to pursue this? Or point me to research that’s already been done on it?