the visual display of qualitative information

If you really want to know what it’s in the criminal complaint filed against Rod Blagojevich, you can read it. But in this 2.0 world, why read when you can visualize? Bill Allison at the Sunlight Foundation’s Real Time Investigations blog uploaded the complaint (via) to a site called Many Eyes, which is the kind of site I wish I’d already known about, and which offers a number of ways to visualize text.

(Unfortunately, wordpress strips out the code that makes it possible to embed these images at their full size and functionality, so I’ve re-sized these images to be larger than what the embed codes were giving me. If you click through, you’ll be able to do all sorts of things, like re-arrange the displays, search for particular words and phrases, count or highlight specific occurrences, and even zoom in on the word tree.)

Let’s start with a Wordle:

blagojevich-complaint-wordle

That gives you an idea of the most important topics/people in the complaint, but it’s more of a bird’s eye perspective. If you want more precision, albeit at the cost of some visual elegance, you could look at a couple of tag clouds:

Here’s a cloud formed on the basis of single words:

blagojevich-complaint-one-word

And here’s a two-word tag cloud:

blagojevich-complaint-two-word

The two-word format does a better job capturing many of the subjects – not just the proper names, but also senate candidate, financial advisor, planning board, campaign contributions, and so on – as well as the alleged activities – Blagojevich spoke, Blagojevich talked, attempted extortion, phone calls. But it also has some pairings that are simply the result of the stylistic conventions of a criminal complaint. For example, the phrase “2008 rod,” which has 53 occurrences, isn’t a phrase in the usual sense – it’s the result of writing out the date of an alleged action (in 2008), followed by Blagojevich’s first name: “…the morning of November 12, 2008, Rod Blagojevich talked to Fundraiser A…”

Finally, the most innovative and analytically interesting visualization is the word tree. Want to know how the Senate candidates appear in the text?:

blagojevich-senate-candidate

Judging by the number of occurrences, Senate candidate 3, who appears to have been identified, is looking pretty good.

Incidentally: a cursory search for profanity doesn’t turn up nearly as much as you’d expect from the news coverage.

1 comment

Comments are closed.