Analysis of content published on the social network Twitter related to Brazilian candidates for the presidency of the republic

Renato Lopo Montalvão Júnior, Victória Felipe Damascena Souza, Raphael Magalhães Hoed

Sentiment analysis, Elections, Data mining, Social network.

This article seeks to verify how related the opinion of Twitter users about the Brazilian candidates for the presidency of the republic is with the polls of voting intentions. Do the positive and negative sentiments on Twitter track the results of the polls of voting intentions for president? In view of this, publications made on Twitter that mention candidates for the presidency of the republic were collected, using python scripts to collect the tweets and to classify them supervised machine learning algorithms via the sklearn library. Therefore, it is intended to verify how compatible the interactions on Twitter are with the results of the presidential polls of the year 2022.

Received: 04 Nov 2022, Receive in revised form: 27 Nov 2022, Accepted: 04 Dec 2022, Available online: 09 Dec 2022

10.22161/ijaers.912.1