Word Cloud Interpretation for Fun and Profit

That last blog posting generated some interest and I’ve been thinking more about the idea of word clouds as a possible adjunct to what we do as qualitative researchers.

Lois Kelly, who is a brilliant word-of-mouth marketer, had a very interesting comment and turned me on to psychologist James Pennebaker’s work. Pennebaker, a psychology professor at the University of Texas, has published some interesting research about his spin on content analysis. His systems counts and classifies the different types of words that people use (such as adverbs, adjectives, verbs, and even pronouns), and also catalogs them by their different emotional implications. He has published, along with co-authors, some interesting work analyzing and comparing different types of written texts and spoken words. He had even started a business from it, selling the LIWC, or Linguistic Inquiry and Word Count software program.

It is intriguing, relevant, high-end, sophisticated stuff, and I am seeing a lot of work in this area as different companies are working hard to develop software that can do sophisticated content analyses of all of the textual information contained online. Companies like netbase, AC Nielsen Buzzmetrics, Cymfony, and Motivequest are all over this emotionally-oriented, passion-and-enthusiasm content coded stuff, and for good reason. It makes good sense.

But when I was speculating about word clouds and their qualitative research usage, I was being a bit more open-ended in my ruminations. I wasn’t thinking about such well-developed systems, ones that provided you with such well though out schemata for analyzing texts.

I was thinking more about doodling. Perhaps even diddling with the doodling.

I was wondering what we could do with the simple word clouds we generate with online engines like wordle.net, which I’ve been playing around with a bit after the Latin American marketing research visionary Pablo Sanchez Kohn set me up with it in one of his online experiments.

I’ve been tooling around with it, and I like it. Why? Well, the people who post it online call it a “toy” and that’s a real insight. It’s just a lot of fun to play with it. It makes decent art, too. The pictures are nice to look at. And to think with, I think. I may frame a couple, or turn them into t-shirts. Why not?

In the play, in that visual appeal, I think there is something deeper that can happen. Because we get some good analysis packed into an at-a-glance form, I think we have opportunities to use that ultimate piece of ‘software,’ the human mind and imagination, to glean insight from these word pictures. We can run poetic comparisons on these scrambled word omelets. And then we can use them to launch creative inquiries into the text. Delving deeper using all our other tools, from content analysis word counts to hermeneutic mindful deep breathing.

A couple of examples today, and then maybe a few in the next post.

Below, I have posted a couple of recent blog pages from Brandthroposophy run through the word cloud generator at wordle.net. I think they’re fairly easy to identify. See if you recognize them.


That one was the recent set of postings having to do with netnographic resources. It is interesting how that big http dominates the frame; the posting is full of links. But it also captures at a glance some central meaning on the page about a community of Internet research scholars coming together to offer a list of new and different places and things to study. The cloud somehow captures all of this for me rather quickly. You get the academic content, the variety of internet forms, as well as the sense of a knowledge network that is building.

Here’s another one. Do you recognize it?


For that one, I took one of my favorite postings, the sad saga about my screws from Costco.com. Can you see at a glance how different these two postings are? This one tells a story about Costco, please Trudy, the screws, the wall, the chair, in March. It’s all somehow squeezed in there. And nice to look at, too.

Unfortunately, I wasn’t able to read in my entire blog, all of the entries from day one, and then construct a single word cloud from it. The software just pulled up one page or entry at a time. That’s too bad, because it would have been exciting to see what the common words have been for the 18 months or so this blog has been in existence. I wonder…

Word cloud: toy, art, or serious research tool. Why should we have to choose?


  1. magda wesolkowska January 28, 2009
  2. magda wesolkowska January 28, 2009
  3. pablosan76 January 29, 2009
  4. Domen January 31, 2009
  5. Bob February 4, 2009

Add Comment