Os padrões espaciais da desigualdade socioterritorial no estado de Goiás, Brasil

Abstract

Socio-territorial inequality can be identified and analyzed based on predictors of social inclusion and exclusion. The objective of this thesis is to map and evaluate the spatial distribution patterns of socio-territorial inequality in the municipalities of the state of Goiás, Brazil. The research is based on the hypothesis that socio-territorial inequality in Goiás can be analyzed by considering social, economic, and environmental predictors. To this end, and through a mixed methodological approach, the study combines bibliometric and systematic reviews, Exploratory Spatial Data Analysis (ESDA), and multivariate statistical modeling using Ordinary Least Squares (OLS) regression and machine learning algorithms (Random Forest and XGBoost). Artificial intelligence was used in the literature review on global studies of socio-territorial inequality, aiming to identify the most commonly used indicators of social exclusion and inclusion. Subsequently, ESDA was applied to map and characterize spatial patterns of inequality, providing a detailed overview of the distribution of socioeconomic factors. The results revealed the existence of significant clusters of regional disparities, with a higher concentration of poverty and exclusion in the northern and northeastern regions of the state. In contrast, areas more integrated into the regional economy, located in the south and southeast, showed better municipal development indicators. This procedure was complemented by statistical modeling, which included OLS regression and machine learning algorithms such as Random Forest and XGBoost, enhancing the accuracy of predictive analyses. This approach allowed for the identification of the main variables influencing socio-territorial inequality, highlighting those related to agricultural production, sanitation, health, and education. The findings indicate that the expansion of agribusiness in Goiás reinforces patterns of spatial selectivity, concentrating income, infrastructure, and services in certain regions while perpetuating historical inequalities in others. Clusters of social inclusion and exclusion were identified, associated with indicators such as illiteracy, precarious sanitation, insufficient SUS hospital beds, and lack of access to basic infrastructure. Although agribusiness is a significant driver of both state and national GDP, its benefits are unevenly distributed, highlighting the need for public policies that integrate economic growth with social justice and balanced territorial development.This thesis contributes to methodological advancements by demonstrating the effectiveness of combining traditional spatial analysis techniques with computational artificial intelligence methods, enabling the identification of complex and nonlinear patterns associated with socio-territorial inequality. From a theoretical standpoint, it reaffirms the importance of considering multiple social, economic, political, and environmental dimensions when analyzing territorial inequalities. In conclusion, the research shows that socio-territorial inequality in Goiás is not an isolated phenomenon but the product of historical and structural processes that require territorialized and integrated public policies. Such policies must be capable of linking the economic development driven by agribusiness with investments in social infrastructure, education, and health. The continuation of studies that combine spatial analyses, statistical modeling, and artificial intelligence is recommended to deepen the understanding of socio-territorial inequalities in other Brazilian states and in similar Latin American contexts.