A New Way to Weigh Politicians' Words


The Democratic presidential debate held two weeks ago at Howard University in Washington, D.C., rounded up a troika of meetings between the eight candidates in which they hashed out an image of their priorities and, collectively, the priorities of the party in this election cycle.

One popular way that blogs condense large amounts of material into a manageable index is with a “tag cloud,” in which all the major themes are mapped out in font sizes proportional to their frequency. The most common subject is the largest, and so forth. So we decided to apply the concept to the debates, creating a tag cloud for each Democratic candidate based on which topical words he or she said most frequently.

U.S. News tech producer James Stevenson and reporter Chris Wilson created the clouds by breaking out the transcripts by candidate and analyzing them with software that counts word frequency and eliminates highly common words, like articles and prepositions, to boil down a text to its themes.

See the tags for the Democratic debates mapped out here.

Clouds for the Republican candidates are forthcoming. Roll over a word to see the number of times it was said.