Análise espaço-temporal da trajetória de carreira de pesquisadores brasileiros

Abstract

The growth of the Brazilian scientific production in recent years is remarkable, which motivates an investigation on the factors, inside and outside the country, which helped shape this wealthy research environment. This work seeks to verify the viability to use curricula vitae data in spatio-temporal analysis, and provide a guide to use geographic visualizations with this data. Towards that end, this article provides a thorough analysis of the education and evolution of the main Brazilian research groups, by analyzing curriculum vitae data of about 6,000 researchers involved in the country's National Institutes of Science and Technology (INCT) initiative. Data on the steps taken by each researcher in her education, from the bachelor's degree to doctorate, including a possible postdoctoral experience, and employment, are extracted from an official curriculum vitae repository. The location and the time at which each career step was extracted from the CVs and are arranged to build spatiotemporal career trajectories, which are then analyzed considering additional factors, including the area of knowledge of the INCTs to which each researcher is associated. The analyses of said data indicate the increasing prevalence of Brazilian institutions in the education of Brazilian scientists, as the number of doctorates obtained abroad is diminishing over time. Postdoctoral stages, on the other hand, are primarily taken in Europe or in the United States. Taking an international postdoctoral position after a full education in Brazil indicates a drive towards seeking higher-level exchange and cooperation with foreign groups in a more advanced career stage. Results also point towards a tendency for Brazilian researchers to seek employment in regions that are close to the institutions at which they received their bachelor's degrees, suggesting low mobility within the country. We also discuss particular characteristics of the geographic visualization created. Maps are good to show patterns and outliers, but a big number of lines may hinder map reading. Filters and opacity levels help map reading, but their presence should be clear for the user. The dimensionability is low, theres a compromise between map readability and how to show attributes. Elements from this study may be used in the definition of public policies for correcting distortions in the researchers education. Moreover, it allows verify the limitations and potential of visualization techniques for exploratory analysis of spatiotemporal curricula data.