A tag cloud displays the 50 words used most often by a single speaker in a speech or interview, but with the font size of each word varied according to the frequency of usage. That is, the larger the word's size, the more often it was used. Common words like "and," "of," "the," and "new" are omitted. Thus, the tag cloud gives a visual depiction of those things the speaker considers most important, at least at the time and/or within the context of the particular speech or interview.
Or debate. Yeah, Mark Blumenthal of Pollster.com has posted tag clouds made by a friend for each of the candidates in the Republican debate last night, as well as for each of the [declared] Democratic candidates in their debate last week. And because the tag cloud device is relatively easy to analyze, it can be an especially useful tool in comparing a number of different speakers, all the more if they are speaking in similar venue.
I thought it would be interesting to draw up a tag cloud of one of General Clarks's speeches. I wanted to use one where he was not limited to or required to cover any particular topic, and the most recent seemed to be his speech to the DNC Winter Meeting on Feb 2, 2007. It is probably not completely fair to compare this to any of the debate clouds, since word content in the latter would be driven to some extent by the questions asked by the debate moderators -- only "some" because most candidates tend to give little pre-prepared speeches to questions they would like to answer, as opposed to those they are actually asked. Nevertheless, I think it does tell us something interesting about the message General Clark sought to impart to his fellow Democrats in this one opportunity to reach the broadest possible party audience, as well as where his emphasis might lie should he decide to declare his own candidacy. And here's what he had to say... there are a few significant words you won't hear from any of the current crop of presidential candidates. Not often enough.
