Text Analysis on User Generated Content: The Rap Lyrics Challenge

Abstract

This research proposes a hybrid approach that uses User Generated Contents as a corpus in order to create a multi-level resource, by combining linguistic, statistics and semantic data. The tool we create overcomes the lack of domain resources and provides a distant reading of one of the most significant cultural and linguistic phenomenon in recent years. We take advantage of techniques from Web Mining, Computational Linguistics and Information Design with the aim to make a large amount of extremely sparse data more accessible and understandable.