Tweeting in the Debate about Catalan Elections

Abstract

The paper introduces a new annotated Spanish and Catalan data set for Sentiment Analysis about the Catalan separatism and the related debate held in social media at the end of 2015. It focuses on the collection of data, where we dealt with the exploitation in the debate of two languages, i.e. Spanish and Catalan, and on the design of the annotation scheme, previously applied in the development of other corpora about political debates, which extends a polarity label set by making available tags for irony and semantic oriented labels. The annotation process is presented and the detected disagreement discussed.