Autor(es): Pérez-López, Roi ; Blanco, Guillermo ; Fdez-Riverola, Florentino ; Lourenço, Anália
Data: 2021
Identificador Persistente: https://hdl.handle.net/1822/74516
Origem: RepositóriUM - Universidade do Minho
Assunto(s): Question and answering systems; Computer software; Bioinformatics; User churning; Social influence; Software development
In this work, different social data mining approaches are used to characterize the user churning and social traits of the Bioinformatics community over the first ten years of Stack Overflow. The proposed workflow consists of a four-step procedure that allows the characterization of users based on the social exchange exhibited by the Bioinformatics community and the Developers communities, notably the Python, Matlab and R programming communities. The motivation is to improve user churning and the quality of social interactions by categorizing user traits and being able to understand how Stack Overflow, and other Stack Exchange databases, may better respond to current user concerns and interests. Therefore, initial user identification was complemented by a second categorization focused on user "genealogy". The goal was to explore the evolving of user interests and, in particular, the migration and swapping of users across different communities throughout the years. Noticeably, a considerable number of Bioinformatics users has moved to the Developers communities. An in-depth exploration of the most popular topics of conversation in Bioinformatics enabled a better understanding of the triggers and contents of these conversations.
