Methodology for analyzing co-authorship networks based on information sources that do not include standardized metadata

Abstract

The research aims to present a Social Network Analysis (SNA) methodology for the Co-authorship Network of a corpus of scientific documents without standardized metadata and specific formats for SNA software, such as those available in databases. One of the specific objectives includes content analysis of the documents. The corpus, with 44 documents published between 2007 and 2023, was collected from the author's Lattes Platform CV, organized into directories and imported into the Mendeley software, where document metadata was filled in and corrected. Subsequently, the metadata was exported in .bib or .ris format for import, configuration and visualization of the Co-authorship Network using the VosViewer software. The Co-authorship Network showed 26 items between authors and co-authors, organized into eight clusters, with 57 links and 118 strong links. Among the strongest relationships are Marcos de Souza and Fernanda Gomes Almeida from Cluster 1, with 9 research projects, highlighting topics such as Topic Modeling, Information Architecture and Research Data Management. The study demonstrates possibilities for ARS from other sources of information, in addition to the metadata provided by databases, broadening the range for specific research, such as identifying the co-authorship network in different areas of knowledge, departments, undergraduate courses or postgraduate programs, leaders of research groups, working groups of events or in an entire university, understanding the relationships between clusters.