We take the raw data streams (e.g. a data mining feed from Twitter) and process them, ourselves.