Estimating Public Opinions Using Twitter Data: The Case of the 2018 Ecuadorian National Referendum and Constitutional Reforms

Public opinion is associated with the study of attitudes, preferences, and beliefs among people toward a specific subject. This study is applicable in many fields, such as politics, to better understand citizens' opinions and to manage them accordingly; for commercial companies, to understand customers' satisfaction; among academic organizations, for research purposes; and in media, for posting articles based on public opinion. The fact that public opinion could be used as a tool in many fields is the core motivation of this work. According to academic literature, one of the most widely used techniques to quantify and interpret public opinion is surveying. Surveying consists of a series of steps to choose a sample population, develop a questionnaire to be answered by the sample population, collect the response data using different methods, and finally interpret and quantify the results. In this work, the Twitter social network is used to collect data in the context of the 2018 Ecuadorian national referendum and constitutional reforms during the three months of political campaigning before the voting. The objective was to analyze public opinions obtained from Twitter and to compare it with the official results presented by the National Council of Elections in Ecuador after the voting. The results presented in this work indicate that Twitter data can be used to extract public opinion with a high level of confidence in sensitive scenarios like elections or voting processes.