Análise da endogamia das coautorias científicas de pesquisadores da UFSCar credenciados em programas de pós-graduação

Abstract

The Lattes Platform has established itself as a fundamental database for the production of Science and Technology indicators in the Brazilian context. Aiming to analyze academic inbreeding in the scientific co-authorships of researchers at the Federal University of São Carlos (UFSCar), this research developed an integrated computational procedure capable of coordinating different technologies for the extraction, treatment, processing, and visualization of bibliometric data. The methodology involved automated data collection from the Lattes Platform, metadata treatment through Python scripts, mapping of scientific co-authorship networks using a Java application, and the generation of visualizations with geolocation features. The analyzed dataset comprises 86,341 co-authorship records in full journal articles up to the year 2024, corresponding to 24,612 unique productions, involving 664 researchers from UFSCar – São Carlos Campus accredited in graduate programs, and 7,322 co-authors with professional addresses registered in the Lattes Curriculum. The results confirm the research hypothesis by showing that the developed procedure made explicit the collaboration relationships that are hidden in the standard interface of the Lattes Platform. Evidence indicates a predominance of non-inbred co-authorships and heterogeneity in inbreeding levels according to the unit of analysis, in addition to showing no linear associations between inbreeding and geographical distance. The research demonstrates the effectiveness of the computational procedure for the empirical diagnosis of academic inbreeding and highlights the potential of visualizations for analyzing the dynamics of scientific collaboration from an architectural and interdisciplinary perspective.