I have been following all of the press coverage of the Obama inauguration this week. There sure has been a lot of it.
One of the most remarkable things I’ve been seeing, repeatedly, is bloggers and different newspapers carrying reports or stories that simply enter the text of Obama’s inaugural speech into a word cloud generating program. There are many such programs available on the Internet. They often also include comparisons of the speech to other word cloud generated inaugural speeches. I’ll include an example here.
Below are the word cloud versions Obama inaugural speech, followed by George W. Bush’s inaugural speech, and Clinton’s inaugural speech.
As you can see, it makes for a quite interesting contrast. It’s almost like poetry, where we are reading the most “important” words into a sort of pattern. “Nation, common people, generation America. Today work, prosperity, spirit, world.” But we can also read in complexity, and word choice. And tell something about structure, and the relationship between words as well. And this allow comparison and contrast. Why is the word America important in Obama’s speech, but not American? Why doesn’t the word “world” appear in George Bush’s speech? And so on…
A nice, flash version, interactive word cloud that actually gives you word counts as you pass your mouse over the cloud is carried in the Toronto Star’s coverage of the Obama Inaugural word cloud. They also have some more historic speeches, like George Washington’s and Abraham Lincoln. You can see that story and those clouds here.
Now, word clouds are becoming popular modes of representation, and I can’t see why we wouldn’t start to think about them as content analysis tool to bring into our qualitative research toolkit. They allow and facilitate a certain kind of graphical representation of verbal, qualitative data that make more straightforward particular kinds of comparison (like this one).
- So, we could compare interviews with matching culture members.
- We could compare different blog pages or newsgroup postings.
- We could compare archival data from the same organization over time.
- We could compare related articles by scholars with different methodological. approaches or from different paradigms.. Or the same scholar over time.
You can see where I’m going with this…right? Somebody needs to write up a rigorous methodology for analyzing word cloud output and putting it into a qualtative data analysis framework. If it’s already been done, I’d love to see it. I’ll have some more related ideas on this soon…and some more original word clouds.