About a month ago, I made mention of the infographic in the NYT about keyword frequencies at both the DNC and RNC. Along the same lines, both Cameron Marlow and Anjo Anjewierden have done similar analyses, this time on the smaller scale of the first Presidential debate.
I don't think that either analysis necessarily presents any surprises, at least for those of us who watched the debates carefully, but that's only because it's easier to grasp a 90-minute debate than it would be were the text(s) more substantial. In other words, relationships and frequencies are easier to gather (intuitively?) in a smaller set of data. What's impressive to me about each of these projects is that they confirm some of my impressions about the debate, and they can tackle much larger-scale texts than I myself would be able to. I hope that both Cameron and Anjo will add the results of the other 2 pres debates and perhaps even the VP one.
And I'll be watching to see how closely those subsequent events follow the patterns that they've laid out for the first...