Big data analytics - Sentiment analysis of Twitter data using clustering techniques

Gathering and analyzing data is becoming an increasingly pertinent task in the highly connected world. Information is archived at alarming rates via social media, and as a result analysts have an expansive landscape of information available to inform future decisions and gain competitive edge. Mining this vast amount of heterogeneous data(big data) classically implies significant computational costs, as well as the need for development of subjective methods by which to comb through the forests of text where the signal to noise ratio is precipitously low. Formally, the blossoming field of Cloudy Data is as of yet only defined at a very high level. Informally, the topography of social media is steep and harrowing without methods by which to grapple with the masses of data in a meaningful (and rapid) way. This paper uses sentiment analysis to process social media data, and explores the use of clustering methods on said sentiment analysis to generate a sparse text corpus in response to a specific subject matter query. This method will be tested on 2012 U.S. presidential election data mined from Twitter.