Now a days, due to recent growth of the Internet and the World Wide Web large volume of data is available everywhere. Various approaches have been proposed in the literature to analyze such large volume data involving large number of dimensions. Real time data streams of large volume can be exploited for knowledge discovery. In the current study, a literature review of the different approaches adapted by various researchers to extract knowledge from Twitter data streams particularly by creating a multi-dimensional architecture comprising of Meta tags with tweets is conducted. These approaches have been reviewed in terms of data retrieval, storage, database design, analysis, visualization abilities and the technologies being used for these components. The study also discusses the technological gaps and proposes some recommendations for the future research.