mrflip's photostream  

Stream through a few million Twitter messages, and emit an edge for every pair of non-keyboard characters occurring within the same message. Simply by observing “often, when humans use 最, they also use 近” you can recreate a map of human languages

Stream through a few million Twitter messages, and emit an edge for every pair of non-keyboard characters occurring within the same message. Simply by observing “often, when humans use 最, they also use 近” you can recreate a map of human languages

Network graph of people on twitter connecting to the topics of Big Data, infochimps or Hadoop