Comments/questions? The easiest place is the original Tweet!

A very quick post here with a work in progress. Below is a word cloud of the text of around 2000 stories I scraped from SoFurry a while back. As is typical with these things, I’ve stripped out the 500 most common English words, but beyond that it’s basically just a word frequency summary (i.e., larger font size indicates the word appears more often) of the most common 100 ‘non-common’ words.

I should stress I haven’t really done anything sophisticated here, and this should be thought of as just a bit of fun rather than ‘statistics’ in any real sense. In particular, the data source is just the story content (via SoFurry’s API) which I ran through an HTML converter I whipped up in R. This means that I haven’t filtered out things like author preambles, so there’s probably a bit of interference there. The dataset comes in at about 13 million words, with over 23,000 ‘cocks’, 13,000 ‘dragons’, and nearly 14,000 ‘insides’.

Word cloud courtesy of this rather funky Python generator!

I have a few ideas for where to go with these data - right now I’m working on comparing the relative frequency of words in furry writing to English in general, which is throwing up some quite entertaining results; watch this space for that. I’m also interested in doing some more ‘statsy’ comparisons of SFW and NSFW content. SoFurry’s API is currently being renovated, so I can’t do too much more for the time being, but figured I’d put this out while it’s fresh in my mind!