Coletando e analisando informações do Facebook com a linguagem R

Abstract

Given that with the emergence of the Internet, people, companies and organizations started to stay connected also in this virtual world. The social networks that appear on the Internet provide the interactivity of its various users. Users of social networks, whether individuals or companies, can express, through comments, different feelings, opinions about what was published. This work proposes to explore the analysis of feel- ings in comments posted on Facebook, which was the social network chosen to deepen this work. In the course of the work will be presented the conception, implementation and evaluation of an approach to trend evaluation in comments of Facebook posts. The proposed approach aims at extracting a set of unstructured data from a Facebook page seeking to identify the trend, positive or negative of what is being discussed. It is also part of the scope of this work to apply the approach in a case study - DataMy Tool, as well as an analysis of the results obtained in the implementation of the approach and possible definition of improvements in it. Through the analysis of the results obtained it was verified that it was possible to identify the polarities in the comments processed. Through the results of the implementation of the approach, it was also possible to verify that the technique of text mining together with the analysis of feelings, used in the treatment of information allows the identification of values in the published messages in an agile way, providing an understanding of the subjective ones exposed, interpreting them and classifying them by their polarity.