I’m still processing all the comments that were left on my blog about mainstreaming OER and Open Education last week. I’ve done a quick summary and as a quick visual overview a wordle of the comments. However I thought I would try and do something a bit more sophisticated and perhaps more enlightening. I’m a bit fan of the visualizations Tony Hirst does using gephi, and have played a bit with it, usually with the help of my former Cetis colleague, David Sherlock. Anyway, I was all set for an afternoon of visualisation delights at the weekend, but gephi didn’t want to play. I’ve since found out it is probably a java/mac issue . . .
However in my search for an alternative I came across Textexture, which will “visualise any text as a network” (using gephi). Below is a screen shot of the network for the blog comments (the embed code doesn’t work for me in wordpress), but if click on the picture you’ll be taken to the interactive network.
I’m not sure how much more this tells me, other than confirming the interconnectedness of ideas. But, it is a pretty simple service to use and we all love a network diagram.
So what is textexture doing? Generating topic models and then graphing the result? (Text analysis is still one of those things I’m largely incognisant of…!;-)
I think so – there is an explanation on the site http://noduslabs.com/cases/textexture-non-linear-reading-machine/. Like most things, I don’t really know what I’m doing but like the instant hit of a network diagram:-)
Hi! I’m the developer of Textexture. For blog comments you could also the new tool I made http://infranodus.com – this may be a bit better suited for what you want to do. E-mail me for the invitation code if you’d like to try it. Also – I’m actively working on it at the moment, so if you want, we can work together on implementing a special functionality for blog comments visualization together.
Reading your post I think it would be great to add the feature where you can actually remove the most connected words, so you can see the context where they appear and in this case the information you get will be more interesting and less predictable.
Thanks Dmitri – that would be fab I’ll email you and try out that other tool. Would love to help developments but be warned I am not that technical and don’t know that much about text analysis😄