Um estudo sobre o mercado brasileiro de ações a partir de dados do twitter

Detalhes bibliográficos
Ano de defesa: 2016
Autor(a) principal: Hugo Silva Santos
Orientador(a): Não Informado pela instituição
Banca de defesa: Não Informado pela instituição
Tipo de documento: Dissertação
Tipo de acesso: Acesso aberto
Idioma: por
Instituição de defesa: Universidade Federal de Minas Gerais
UFMG
Programa de Pós-Graduação: Não Informado pela instituição
Departamento: Não Informado pela instituição
País: Não Informado pela instituição
Palavras-chave em Português:
Web
Link de acesso: http://hdl.handle.net/1843/ESBF-AFWLG6
Resumo: In this dissertation, we present a study of the Brazilian stock exchange market based on a large characterization and analysis of Twitter data. In our analysis, we show that events and news about the stock market are capable of generating peaks of postings by Twitter users and that the frequency of posts follows the starting of the exchange trading day and maintains for about three hours after the stock market closing hour. Moreover, based on a survey conducted with a specific niche of Twitter users, we have been able to estimate that 0.5\% of those users have some knowledge of the Brazilian stock market and are mostly individual investors interested in posting and consuming news about this market, having 45\% of them used Twitter as a source for investment decisions. Finally, we have observed that the total number of orders and the financial volume are positively correlated for 66\% of the stocks mentioned on Twitter, whereas the oscillation and maximum oscillation dimensions present no correlation. In particular, we have found that 82\% of the URLs shared in the network are news and that these links are highly replicated, being only 18.2\% of them unique. In addition, we have measuredtheir exposure time on Twitter and shown that it varies from 0,9 to 43,4 hours. In general, this study provides a base of behavioral patterns, both related to the Twitter and the stock exchange, to develop social indicators to support decision.