Coleta e seleção de perfis de usuários de redes sociais para contribuição voluntária

Abstract

Nowadays, millions of people have been interested on resources and tools that allow citizens to publish contents on the Internet. Crowdsourcing initiatives aim to enroll those data using citizens as human collaborators. However, those initiatives does not guarantee that citizens keep contributing on it. Thus, the contributions became flawed and biased over time. In that context, volunteers could be found on social networks based on their thematic preferences. Thereby, the main goal of this research is to identify and select users from social networks, and find volunteers who might work on crowdsourcing applications based on their interests. For that, the methodology consists in performing an interactive process for increasing the collecting data, and thus increasing the content collected. A case study was developed to validate the method based on an urban context identifying initial keywords, which allowed to come up with a refinement process. The results have shown that most of the collected data was not still related to a specific interest. However, the last iteration of the filtering process was better than the initial one. Nonetheless, it is not possible to claim that users posting comments related to a specific interest or subject implies that they have a personal preference for it. Wherefore, adaptations can be done on the filtering process in order to confirm the users preferences.