[Please note that the text links in this piece lead to the live visualizations which are both larger and interactive]

I’ve been playing around with Many Eyes (if you are not familiar with it, it is an easy-to-use data visualization suite; you might call it the flickr of data sets and visualizations) steadily since early January when I attended an interesting HICSS session on how much “unintended” use the system was seeing. I’d originally tried it out almost exactly a year ago, but the service has grown by leaps and bounds. It certainly has great potential for educators and teaching, and is already seeing some use there.

I don’t work with a lot of numeric data and most of that is not something I can make public (a good decision on the part of Many Eyes is that all data sets– as well as the visualizations– are completely public and open for others to build visualizations on), though I have tested various sets to try out the wide array of visualization types. So I have focused primarily on the free text visualizations: tag clouds and word trees.

As an example, here are the one- and two- word tag clouds I made of all my blog posts from my Cosmopoetica Blog:

And some stills of working down a word tree based on the same data:

It’s interesting to compare those to the same visualizations for this blog…

The process of creating a data set is simple: you simply upload the data through a web form through copy and paste from a text file or a spreadsheet. The text area is pretty smart about interpreting the data and doing the right thing and the text filters take out common stop words (articles and many prepositions). The only thing I did to all of these text files was a simple search and replace to delete all numbers and a couple of meaningless stop words like the title of the blog.

Following up on a good idea by George Siemens relayed through a Tweet by Alan Levine I also did some visualizations of the last five years of Horizon reports. I did each individually (2004, 2005, 2006, 2007, 2008) and then one for the combined texts of all years. Here is a still of the combined cloud for all reports, single and double:

It’s really quite illuminating to compare the years. Note how objects disappear in favor of social and social networks and then data and information come in late (no doubt the Web 2.0 meme). And is there any better sign that the NMC report is on the right track than the two-word tag cloud? Augmented reality, creative expression, collective intelligence, social networking, virtual worlds– it’s the bingo card we all want to fill out.

[Incidentally, it took far longer to click around to each set and create the visualizations than it did to create the data sets... all I had to do for those was save the PDF files as text and again, replace numbers and three common words: horizon, report and reports, and upload.]

Another interesting feature of the Many Eyes system is that any particular view of a visualization that has been manipulated can be integrated as a snapshot with comments. For instance, there is a visualization of book publishing trends, below which are comments, many linked to specific views of the data. Good stuff.