Alex Steer

Data science for better marketing / about / archive

How dead is everything?

271 words

Taking my cue from Richard's post, I thought it would be fun to see the extent to which the blogosphere (as read by Blogpulse) has decided that things are dead over the last six months.

A warning first: like most untagged corpus searches, this probably returns quite a lot of noise, especially since Blogpulse tends to ignore stopwords. So don't take it as gospel.

Let's look at how dead email, print and Twitter have been recently. This search looks for the form '[medium] is dead' (which means '[medium] [any stopword] dead' to Blogpulse).

Image lost in database transfer - sorry!

Sorry, print, it's not looking good.

This is the same set of media queried as 'death of [medium]' (i.e. 'death [any stopword] [medium]').

Image lost in database transfer - sorry!

This version is hit quite badly by the death of Michael Jackson, which seems to have occasioned a lot of blog posts about the widespread reporting of his 'death on Twitter', which come into our results because 'on' is a stopword, so indistinguishable from 'of' to Blogpulse. This is a lot less reliable than the previous graph, then, but worth including as a warning to people not to take search results at face value.

And finally, given that Blogpulse is a corpus of blogs, this has to be my favourite:

Image lost in database transfer - sorry!

Does this suggest blogging is a winter sport?

# Alex Steer (03/12/2009)