Sentiment Analysis on Twitter to Measure the Perception of Taxation in Colombia

Abstract

Twitter is a social microblogging tool in which users can publish textual content up to 280 characters. Since its birth in 2006, Twitter has become one of the most widely used platforms for sharing current content, such as news or events in real-time, which also allows you to briefly and concisely comment on virtually any topic. This research presents the analysis of public opinion feelings about taxes on digital platforms to support the decision making of the National Taxes and Customs Office (DIAN), and the Ministry of Finance, for which an extraction and compilation of tweet data is developed, a preprocessing of information in order to obtain a corpus in Spanish that can be classified into positive, negative and neutral categories; then learning algorithms are tested, which allow training models to subsequently classify new collected tweets.