I enjoyed the concept of Gephi, though the operationalization of it was tricky. I enjoyed seeing how the sea of words and cells from our timeline spreadsheet was turned into a map of sorts through the graphing tools, though I also found some of the limitations amusing (my program kept mapping TS and Eliot separately - no surprise, they had many shared edges!). It would be neat to see this program cleaned up and made more user-friendly.
I see one of Gephi's great strengths lying in the way it seems to help overcome some of the difficulties with reading magazines through their online PDF or screenshot instantiations. In Dr. Latham's "Unpacking My Digital Library" piece, he discusses how the common approach to reading a magazine is to flip through an issue, stopping at various articles/scriptons, maybe going through the piece a few times with different sequences, but not to do a linear reading progression from start to finish. The presentation of digitized versions of magazines we've been looking at lately has created an environment that tends to constrain the reader to start-to-finish reading; it's hard to flip through a PDF the way you can flip through a magazine. In contrast, Gephi makes it easy to hover over various nodes and look at their connections at will. In this way, I think Gephi helps restore in the digital realm an important element of and approach to reading magazines.