Modelagem de padrões espaço-temporais: estudo de caso de crimes contra a dignidade sexual no Estado do Paraná-Brasil

Abstract

Sexual violence is a social phenomenon of global proportions, affecting children, young people and adults, mostly female, both in public and private environments, having as main aggressor men from the victim's family environment. The real dimensions of this problem are unknown due to the social stigma surrounding the subject, bringing low notification of crimes to the authorities. For this reason, studies that use a statistical approach around this theme are vital for the consolidation of tools to support society. Thus, this study is important to deal with real data on crimes against sexual dignity by Integrated Areas of Public Security in the state of Paraná, for the years 2012 to 2019, using spatial-temporal mapping and modeling, enabling an accurate interpretation of the phenomenon under study. The modeling approach used is Bayesian, as this is a social epidemiological study, it uses the model with poisson distribution and autoregressive conditional-type priors to capture the spatial effect and smoothing of the temporal component in a linear fashion, based on parameter estimation through INLA. With the spatiotemporal modeling obtained, it was possible to show the pattern of incidence of crimes against sexual dignity in Paraná for each AISP, from 2012 to 2019, through thematic maps and temporal evolution graphs, using the relative risk metric. The results showed a standard temporal space, where it was clear that there is an increase in the number of notifications of cases of sexual violence in the state, with the exception of one AISP. It is concluded that the INLA approach proves to be a powerful tool in spatiotemporal analyses, especially when the treatment of information is about delimited and specific areas. This work is important because it clearly brings out the reality of violence in the state of Paraná and allows for specific and diversified interventionist proportions that can be applied in certain areas according to their particularities and disparities.