Tracking Targeted bits in Twitter

This work deals with the definition, extraction and graphical analysis of conversations extracted from Twitter. Twit- ter is a popular social media platform wherein members submit (mostly) public content. This content is generally in the form of tweets, short messages up to 140 characters in length. This work uses an advanced crawling technique to mine conversations from Twitter and then subjects it to various analyses. This type of conversation is an excellent tool for studying human interactions. The analyses in this work include topological, temporal and probabilistic tests of tweet diffusion. Furthermore we investigate methods for classifying conversation graphs and for identifying user roles based on topological centrality.