Spatio-Temporal Trend Analysis of the Brazilian Elections Based on Twitter Data

Text classification techniques and sentiment analysis can be applied to understand and predict the behavior of users by exploiting the massive amount of data available on social networks. In this context, trend analysis tools based on supervised machine learning are crucial. In this work, a framework for spatio-temporal trend analysis of Brazilian presidential election trends based on Twitter data is proposed. Experimental results show that the proposed framework presents good effectiveness in predicting election results as well as providing tweet author's geolocation and tweet timestamp. According to our results the spatio trend analysis applying our framework via SVM on the Twitter data returns an accuracy close to 90% when the Support Vector Machine (SVM) algortihm is applied for sentiment classification.