Sentiment Analysis Applied to the Popularity Level of the Ecuadorian Political Leader Rafael Correa

Social networks have become indispensable tools for the interaction of political figures and their constituents. For which research has been conducted on the analysis of feelings in social networks with techniques for processing natural language. This article's contribution sentiment analysis using a tool for NLP adapted to the variation of Spanish used in Ecuador, taking advantage of the fact that most of the literature has focused on the English language, while adaptations for languages, like Spanish, are minimal and still in process because of the inherent complexity of language. In Ecuador, the social network Twitter is one of the main means of direct interaction between political figures and the population. Similarly, due to the complexity of a study that reflects the feelings in Spanish for idioms from each region, it gives us a great opportunity to study the relationship between a candidate's level of acceptance on Twitter and the electoral results. This study makes an adaptation of the Stanford NLP tool to the regional language of Ecuador. Regional expressions, idioms, messages with discordant meanings, abbreviations, and other characteristics of the Spanish language were reviewed. Subsequently, the feelings of the messages sent to former President Rafael Correa were analyzed, contrasting them with the results of the pro-government candidate in the 2017 elections. Additionally, we find that the feelings towards the top Ecuadorian political leader did not influence the electoral results of his party.