It seems that I’ve reached the final stages of my dissertation research (The writing stands at about 45%). Last weekend (yes, we were working late on Friday) my lovely collaborator and myself finally sat down to outline the data mining process we’re going to carry out and what we’re going to do with the data. The problem is not going to be actually getting the data, which seems pretty straightforward and it should *only* take about 28 hours in one computer, 14 in two, 7 in four, and so and so on. The problem is how we are going to process and visualize what is promising to be a network of about 70 000 nodes and 100 000 edges without burning up all of the Cultureplex computer infrastructure and, more importantly, so that it tells us something. We are working on that.
During the process I’ve started outlining my hypotheses for the
experimental part of the research, based on the data I already have. And this is what it is, so far.
The structure of the Orsai network based on data gathered so far considers year 1 and 2 of the print publication, ie. issues 1 to 10 and all blog entries from September 2010 to Jan 2013 and is organized according to this schema:
‘author’ (in blue) writes a ‘piece’ (in pink) that is published in a ‘medium’ (in red). Fig 1.
Up to here, the network is pretty shallow and straightforward to follow. From here, it is evident that both editors are also the most active ‘authors’. In a different visualization, other authors start to stand out as the recurrent collaborators, while the large majority remain one time contributors to Orsai.
The hypohtesis pretty much sustained by this is that Orsai can be read, largely, as a ‘one-pen’ narrative, at least as individual as we can expect in these days and in these media. Hypothesis to explore in the near future: Names will fluctuate more as I include the readers – as the commentators of the project. Our estimate so fat is that on average, each of the almost 500 ‘commentable’ pieces has 100 comments.
Now, not only is Orsai more stable in terms of authorship than it would seem at first, but also, although there are many genres within the different media: profile, narrative chronicle, popular arts, comic, fiction, folletín, etc., the self-referential genres (leaks, editorials, sobremesas and entradas, etc.) amount to over 13% of the total of the pieces. This confirms my early intuition that Orsai is largely about Orsai, and that readers expect to read about Orsai in Orsai. Hopefully you’re still following. (Fig. 3)
From here, I started wondering where all the self-referential pieces were coming from: From the editors is the first response, and I’m right about it. But, I also found out that they are pretty much scattered proportionally in all the five media, which again confirms my earlier hypotheses about how Orsai keeps its cohesion among all its different manifestations. (Fig 4). This is an interesting insight up to know. What will be even more interesting is to see what other pieces can be marked as self-referential from the natural language analysis that’s going to come. What other authors are commenting on Orsai as they publish in it? And still even more interesting, to what extent are readers commenting on Orsai itself? Is the graph going to end up all died in pink? Maybe
So what’s next. After the data mining super session. This little (800+ nodes) graph will become a monster to include all the readers, all their comments and the super intricate relationships between them. Basically, Who comments on what? Who replies to whom? How much? etc. And these are some of my hypotheses:
1. During the first year of Orsai more readers were making fewer comments but amounting to a larger volume (many comments-many readers). This tendency has likely reverted, and for the second year, or for the second half of the second year, fewer readers have amounted for the majority of comments (many comments-fewer readers).
2. Self-referentiality, ie. talk about Orsai itself has been a constant through the two years of publication. The terms to refer to it, however, have changed along with the project(blog, magazine, bar, editorial, etc). What has been the development? I want to suggest it has increasingly been referred to in more concrete terms and notions that the readers can more closely relate to. This question will be solved with NL analysis during the latest stages.
3. A tight community with very clear ‘code’ markers of interaction (like the PRI game at the beginning of each comment thread) has taken shape. What the development of that community is, I can’t tell yet.
4. Because of place-specific ‘props’ (pizzeria, bar) in addition to distribution issues, the readership of Orsai has become more local than it was in its earlier days. Answers to this will also show up after the data mining session.
Does anybody want to put some money on any of them?
Also, for more, you can visit my poster on Research Day to be held on March 25. Western University, Great Hall. I will be there, and there will be cookies and hummus.