It’s good to see that important, important work is being done in this area of semantic analysis: “geek” vs “nerd” at the blog slackpropagation. I think the author (with some commenters) has a point in realizing that using Twitter as his data source possibly limits and no doubt skews his results. As he mentions, it would require work with the Ngram corpus or suchlike to be more thorough. Nevertheless, I appreciate the attention to detail in his work.
