Análise socioespacial da COVID-19 no estado de Minas Gerais: uma abordagem por geotecnologias

Abstract

COVID-19, first identified in South America in São Paulo in February 2020, quickly spread throughout Brazil, impacting the population unequally, especially the more vulnerable communities. Minas Gerais, in particular, stood out as the state with the second-highest number of cases and the third-highest in deaths, revealing significant socio-spatial heterogeneity. In this context, the aim of this study was to analyze the socio-spatial patterns of COVID-19 distribution. The methodological framework included: i) data collection on COVID-19 (cases and deaths by municipality) and organization into spreadsheets; ii) analysis of geospatial clusters using Geoda software; iii) obtaining socioeconomic variables for modeling using a machine learning algorithm (Random Forest) to explain the socio-spatial profile of COVID-19 in Minas Gerais. During the analyzed period (March 2020 to July 2023), Minas Gerais recorded over 4.2 million cases and about 65,740 deaths, with a fatality rate of 1.56%. All municipalities reported cases, and only six did not report deaths. Using the Moran's Index (Cases = 0.424820, Deaths = 0.399077), two contrasting cluster patterns were observed for the state, in a north/south direction. In the northern part of the state, low-low clusters were identified, that is, this group of municipalities had low numbers of cases and deaths. Conversely, the Intermediate regions of Uberlândia and Uberaba showed high-high cluster patterns, indicating groups of municipalities with high incidence and mortality rates in Minas Gerais. The Random Forest was trained with 16 socioeconomic variables to explain the cases and deaths. The model showed that 84% of confirmed cases were explained by variables such as the population aged 65 and over, total population, availability of beds, HDI, and ventilators. Meanwhile, 92% of deaths were explained by total population and the population aged 65 and over. In this context, it was found that the distribution of COVID-19 cases and deaths in Minas Gerais is associated with demographic and socioeconomic factors. The findings of this thesis are useful for public policies and prevention measures, with the potential to be applied in more detailed analyses at the municipal and neighborhood levels, facilitating the understanding and control of COVID 19 and other diseases.