Space-temporal Analysis of the Homicides Frequency in the Natal-RN districts from 2011
to 2021
Space-Temporal Analysis; Intentional Lethal Violent Crimes; Bayesian Models;
CARBayesST.
Areal Units comprises information, quantities, or indicators that represent
contiguous areas of a specific region, such as countries, cities, among others. They are subjects
of study in various fields such as Geopolitics and Cartography, Urban Planning, Environment,
Public Administration, and Health. Several software programs have been developed to assist in
processing these types of data, such as CARBayesST, developed for the R language. It utilizes
space-time structured Conditional Autoregressive Priors (CAR) within a Bayesian Hierarchical
framework and employs Monte Carlo simulations via Markov Chains. The phenomenon of
violence has existed since the dawn of civilizations. It is the result of the complex interaction of
various elements, exacerbated by segregation and exclusion stemming from the urbanization
process. The term "crime" is used to describe the violation of the laws that govern a society,
which are regulated by a justice system backed by the State. The city of Natal, the capital of
Rio Grande do Norte, transitioned from being the least violent city in the country in 2003 to
the most violent in 2017, with a rate of 102.56 homicides per 100,000 inhabitants. In light of
this, an analysis was conducted to assess the impacts on the number of occurrences of
Intentional Lethal Violent Crimes (CVLIs), which encompass deliberate homicides, bodily injury
followed by death, fatal robberies, feminicides, and other offenses leading to loss of life
(excluding deaths resulting from state security actions) in the neighborhoods of Natal from
2011 to 2021, utilizing the CARBayesST package. The assessment of the best fit and
convergence was conducted based on the criteria proposed by Geweke et al. (1991),
Spiegelhalter et al. (2002), Watanabe and Opper (2010), and Congdon (2005). As covariates,
information regarding public lighting expenditure in megawatt-hours (Mwh), the number of
public facilities such as parks, police stations, and schools, the average nominal income in
terms of minimum wages, the illiteracy rate, the number of enterprises, and the percentage of
urban pavement were selected. This information was extracted from the Municipal Secretariat
of Environment and Urbanism and the 2010 Census. Additionally, population estimates and
population density were obtained from the Worldpop Project (2013), which utilizes census and
geospatial data, along with satellite imagery, statistically modeling them to produce
population distribution estimates in various global areas. Since the study period encompasses
the COVID-19 pandemic, based on Freitas (2021), the addition of a dummy covariate for the
pandemic years was considered necessary.