Although I feel comfortable with what social network analysis does at a conceptual level, the thought of performing my own experiment is a bit overwhelming.  I considered creating a visualization of the various character interactions in a single chapter in Ulysses, but this task proved to be far more difficult than I had imagined. At the moment I am finding a case example more helpful, and so I checked out Amanda Visconti's (PhD candidate, University of Maryland) blog as it pertains to her collaborative Bloomsday 2013 mapping project.

http://www.literaturegeek.com/bloomsday2013results/

Perhaps one of the reasons I found the process of logging various character interactions so difficult was because it required that I define what constitutes an "interaction;" moreover, it requires that I code different kinds of interactions.

Here is Amanda's visualization of the "Wandering Rocks" chapter of Ulysses:

The thickness of the lines corresponds to how each interaction was weighted: dialogue is considered a more intimate interaction ( or "mutual interaction" in Amanda's terms) and is therefore weighted more heavily than "lighter" interactions such as a character thinking of another character, or gesturing to them in the street (or "directional interactions").  The scale of these projects sometimes means they are better as collaborative efforts, but Amanda's Bloomsday 2013 project illustrates how such tasks require an great deal of planning and strategizing. Amanda had to address the fact collaborators were using different editions, and she had to specify to her colleagues what constituted an "interaction" and how different kinds should be coded in the spreadsheet.

In class two we discussed how marking up a text may be quite difficult if the text doesn't clearly delineate what words are spoken or thought, and this poses a similar problem when trying to code different interactions. Amanda notes that determining when interactions begin and end will necessarily be a bit subjective.  I found Amanda's coding criteria quite helpful in terms of understanding how to break down the interactions, but she admits that it needs to be refined so there is less ambiguity about directionality (especially in codes 4-7) and also a clearer sense of how many times an interaction should count (because there is potential for overlapping, "X thinks about Y" while "X talks to Y" etc.):

Lastly, Amanda generated a map of the "Wandering Rocks" via Gephi:

I wasn't entirely surprised to see that Father Conmee was one of the core nodes in this chapter, but I didn't expect Simon Dedalus, or for that matter the (seemingly) peripheral figure of the one-legged sailor, to be such central figures (though not the most important characters by any means, as this chapter doesn't have a dominant or crucial figure). This visualization emphasizes how Ulysses, as a celebration of the quotidian, reminds us that the majority of our daily interactions occur with strangers or mere acquaintances.  I also found it interesting that this map (more so than the previous visualization) suggests that the urban activity is built around an "empty center" of sorts, and is thus able to capture what the text does at a formal level through its use of a third person limited omniscient narrator who weaves the stories together but has no presence of his or her own.