Public opinion in Twitter. Analysis during the electoral race in the Colombian capital city

Abstract In Colombia, electoral polls have been losing credibility, so rigorous monitoring and analysis of social networks during an electoral campaign could be a way of predicting election results. Which indicators of public opinion on social networks determine a candidate´s triumph? The research that comprises this article traced on Twitter some key theoretical components of public opinion that could accurately anticipate the winner during the electoral process in Bogotá. Using ATLAS.Ti, our research efforts coded posts on candidates’ pages to generate data for our analysis. It was possible to identify seven (7) public opinion indicators that in an aggregated way coincide with electoral results. Likes on Facebook or Twitter were most effective in anticipating the candidate elect. The originality of this research is based on its methodology for analyzing content posted on Twitter and the innovation of including feelings and moods as important aspects of attitudes, the deepest component of public opinion. Our research results validate public opinion theory and its inner components. Survey agencies must adjust their methods to preserve credibility.