Tweetalyser : a Twitter based data mining system with recommendation capabilities

Abstract

Since its inception, Internet has entered most peoples' lives. The deeper the study of the medium, the more uses are found for the Internet. Due to the constantly growing variety of content-rich multitasking handheld devices and the advent of Web 2.0, the user with the most basic computer skills, can post anything anywhere. A vast a rich database is being amassed thanks to the burgeoning popularity registered in the sphere of social network sites. One of the challenges posed by this database however, is not only finding information, but finding information that is relevant to the user. Both academic researchers and business alike appreciate the huge potential that this data possesses, especially in the marketing field. Twitter, one of the most popular social networking sites, provides a treasure trove of untapped resource which researchers can mine for information. Tweetalyser processes and stores the free information available in user tweets and uses it as a base for item recommendation in a scientific approach. The system proves that it is easy to tap into this data mine using low footprint social network APls and third party technologies, in conjunction with the appropriate scientific algorithms. When at its full potential, Tweetalyser is able to provide the researcher with useful information about a user which, when put in the right context can uncover true knowledge. From a social networking point of view, Tweetalyser not only has the potential of recommending and thus increasing attendance to various events, but also attracts newcomers who in turn are able to crowdsource useful data elements to construct a truly formidable knowledge system.