Twitter StoryWanglerAt the Computational Story Lab, since 2008 we have collected a random 10% of all public messages using Twitter’s Decahose API. Overall, our collection comprises roughly 150 billion messages requiring 100TB of storage. Inspired by the Google n-grams project which smashes books into bits, we’ve meticulously parsed these tweets into daily frequencies of words, 2-word phrases, and three word expressions. Scientists call these different types “n-grams” for n = 1, 2, 3.