Usando mapeamento sistemático para análise de grupos de pesquisa em engenharia de software.

Abstract

The publications made by groups, laboratories or research communities (articles, dissertations, technical reports) have been increasing in the area of Software Engineering (ES). The researchers developed by these groups are critical for their consolidation and reference in the scientific community. In this context, systematic mapping is an experimental method that presents a process of analysis of data collected from primary studies. The results of this process are presented through visual maps, which facilitates the researchers' understanding of a research area. This method, used in ES studies, can also be used to analyze publications from a laboratory, group or research community. This work presents the MASG method - Method for Systematic Analysis of Research Groups - that allows to characterize groups, laboratories or research communities, with the purpose of assisting its leaders and researchers in decision making as to the direction of the new researches to be carried out, as well as identify research strategies used by the group. To evaluate the method, a case study was carried out with the Software Engineering Laboratory (LES) of the Federal University of Bahia (UFBA), focusing on Software Repository Mining (MSR). As a result it was possible to observe that the proposed method allows to characterize researches of a group, to identify its profile and research strategies. It also allows leaders and researchers to identify new research opportunities to be explored by the group through the analysis and comparison of graphs and visual maps, as well as to gain an understanding of the position of the group in relation to the research being carried out in an area of interest. Individual comparisons of the researchers were also performed. This process characterized and positioned the researchers in relation to the group and area analyzed, and provided the same subsidies for decision making regarding their future research.