Here's a podcast from that list-- -- 17 minutes of looking at big data analysis of language from Kieran Snyder from Textio.
It turns out that Kickstarters that get funded are more likely to have a lot of text and (depressingly) a lot of different fonts.
And (their main product analyses job offers) that "synergy" turns job applicants off, and women are less likely to want jobs that advertise for gurus or ninjas. You can run other sorts of data through their job offer analyzer-- you might find something interesting.
I'm worried about what happens if big data is commonly used for language-- there could be a lot of tiresome convergence.
This has a text version some of what was in the interview.
This entry was posted at http://nancylebov.dreamwidth.org/1071374.html. Comments are welcome here or there. comments so far on that entry.