This tag cloud was generated from the 2008 State of the Union address. The words which come up in the very largest font are no surprise. Saddening to see “liberty, freedom, progress, peace” quite so much smaller than “terrorist, security, fighting” though. The relative size of “tax” compared to “health” or “schools” too. “Plan” is hardly there at all.
Via Matt Cordell from On Day One, a site set up to brainstorm goals for the next president of the USA, and especially on some actions to get started right from the first day in office.
The tag cloud was originally generated by Jason Griffey, who also did one for the SOTU last year. Last year “health” and “insurance” were huge themes, this year hardly a peep.
A few major differences: Congress is mentioned a lot more this year, while health and oil don’t show up at all. This year’s address looks more active…instead of “fight” we get “fighting”. Overall, the themes are still easily picked out: “terrorists” still play a major part in the speech, and we get “empower” and “hope” and “trust” as themes.
The Net-Savvy Executive picked up on Jason’s tag cloud last year and wrote this post on Visual Text Analysis, showing how other tools can also be used to analyse the SOTU. Apparently lots of language analysts love comparing the SOTU from year to year.
Categories: language, Politics
Today in nodding furiously: what do you mean when you say you’re entitled to your opinion? 
Things that unexpectedly remind one of Mark Latham 
About #ditchthewitch 
Feminism – we’re doin it rong #1568454876 
Hoyden About Town 
Interesting tool. Are there any widgets or gizmos to do the same sort of analysis on other sources, like blogs?
Yep, what Jason used to generate the tags above.
http://tagcrowd.com/
I had a fiddle with it, but it seemed like it could only analyse a single webpage or an uploaded/pasted file, not a whole website/blog archive. Yes?
I believe so.
tig, if you are not already doing so, Jon Stewart is the best filter for American politics. Even without writers his take on the SOTU address was brilliant.
Interesting to compare the textual analysis to The Daily Show. ‘Trust’ is a key word that both picked up.
Shaun’s last blog post..Some honking great rock?n’roll
When analyzing texts, ontological types create frames of two and/or three words, and make frequency lists of combinations (not permutations) of word roots in those frames. Are there any tag-cloud creators that look at texts using this more sophisticated view?
Dave Bath’s last blog post..Hypocrisy from Vic shadow agricultural minister