Estatísticas oficiais e capitalismo de plataforma: a transição para um regime de dataficação no Brasil

• Ano de defesa: 2024
• Autor(a) principal: D’Alva, Oscar Arruda
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Estatísticas oficiais; Plataformas; Big data; Ciência de dados; Official statistics; Platforms; Data science; CNPQ::CIENCIAS HUMANAS::SOCIOLOGIA
• Link de acesso: http://repositorio.ufc.br/handle/riufc/77730

With this work we seek to understand to what extent contemporary processes of datafication and the introduction of big data and data science at the Brazilian Institute of Geography and Statistics (IBGE) reconfigure the field of production and distribution of official statistics in Brazil. Using tools from Bourdieu's genetic structuralism and from the sociology of quantification, we conducted a comparative investigation that addresses the implementation of the Regional Big Data Hub for Latin America at IBGE, in light of the historical processes of structuring the statistical field and transition of statistical regimes, in the contexts of the Global North and Brazil. The research consisted of the analysis of secondary data, compared with primary data, obtained through interviews and non-participant observation. The topic is relevant to the extent that, in the context of Platform Capitalism, the incidence of Big Tech in the transnational statistical field implies risks to the sovereignty of countries in the Global South. Our hypothesis is that the introduction of new data sources and methods in official statistics implies an intersection between the statistical field, traditionally linked to the State, and the algorithmic field, associated to private corporations. This engagement leads to a transition to a new datafication regime. Our thesis is that this transition is characterized by two obstacles. An ontological obstacle, related to the political economy of data, tensioned between the concepts of public good and commodity. And an epistemological obstacle, related to the methods, tensioned between the frequentist and bayesian interpretations of statistics. These obstacles lead to a double movement in the statistical field. Efforts to create data markets for official statistics and to bread data scientists at National Statistics Offices are opposed to initiatives aimed at regulating, protecting and accessing data as public goods, ensuring their incorporation with the preservation of methods and traditional statistical sources, under the governance of State statisticians. In the case of Brazil and countries in the Global South, the relevance of national counter-movements, protective of the public nature of official statistics and data sovereignty, stands out in determining the degree of autonomy or dependence that their statistical systems may assume in the transition to a new datafication regime.