Project

BUZZwatch: Tracking Themes Within Text

BUZZwatch distills and tracks trends, themes, and topics within collections of texts across time (e.g. Internet discussions, newspaper archives, and Web pages). This system works by combining new natural language processing techniques with text analysis and retrieval methods, and the novel application of statistical time series analysis. Applications for the system include data mining and visualization of topic dynamics within texts and Internet chatrooms, as well as improved retrieval for Web search engines.