Geographic and disciplinary distribution of the Brazilian's PHD community: patterns of the scientific collaboration structure

The study of national academic characteristics is an imperative task for the understanding of national scientific production and the creation of effective science policy. Using a dataset of more than 3.2 million Brazilian curricula, we explore the academic community of PhDs working in Brazil in order to identify characteristics of the whole national network and in the knowledge area level. We used metrics from social network analysis and text mining techniques, as well as the patterns of collaboration between areas and the regional distribution of PhDs. The results show different general characteristics of the PhDs working in each Brazilian state and knowledge area, according to the social and economic characteristics of each of the five Brazilian regions. Different interaction profiles were described, like a less connected network in Linguistics, Letter, and Arts, in which each researcher is related, on average, to less than three other PhDs; on the opposite side, Agricultural Sciences each researcher is related, on average, to more than nine other PhDs of the network. It is clear that besides the capital and one or other major city, the Northeast Region is devoid of PhDs, a situation that is particularly problematic for the most destitute region of Brazil.