Mapping comment threads
For some time now the project has had a side-interest in commenting behaviour on on-line news articles. Who talks to whom? What sorts of dynamics emerge? What determines the comment structure: the order in which comments are displayed? Anonymous commenting? At some point in the hopefully not too distant future, Alfred, John, and I will publish a large-scale analysis of commenting patterns on the Huffington Post.
Before moving on to analyse large datasets, though, it’s worth convincing ourselves we can model interactions on individual texts. To that end I wrote some code that will visualise a comment-tree when stored in the translatable JSON format. Conveniently, this is precisely the format used by the Daily Mail (And Facebook. And many others beside).
What does all this have to do with conspiracy theories? More conveniently still, the Daily Mail
recently published an article citing a talk Richard gave about conspiracy theories.
The eagle-eyed reader will note the article is now more than a year old, and indeed, this is an experiment from June last year, which never made it off our local server. Recently I’ve been getting the digital side of the project going, not least on the technical front: we’ve moved the blog to WordPress, and I’ve also finally made some use of the web-space kindly given to us by the University. The upshot of it all is that we can now both host and present interesting visualizations such as the one below.
The visualisation is intended to document how people interact, whether comments are well or poorly received, and also how the conversation changes over time. I tried to pack all the information contained in the data into the frame, so go ahead and try clicking on things. There’s quite a detailed howto – in the menu bar on top – which explains some of the technical detail.
Click ‘GO!’ to launch the visualisation.