The death of George Floyd in the United States made a strong mark in the fight for human rights and against racism. The shocking video published on the Internet showing how events happened brought people around the world to the streets to demand the end of racism and appropriate police conducts. With that in mind I decided to find more data on how Brazil is related to these topics and found this publication from BBC, showing a record in killings by police officers in the state of Rio de Janeiro in 2019.

Rio de Janeiro is famous for being one of the most beautiful places of the world. Unfortunately it is also known as a really violent city, with high criminality rates and public safety issues.

Rio de Janeiro — Photo by Pixabay from Pexels

In BBC’s article it is possible to read about a “policy of confrontation” that was implemented by the safety agencies in the past to help in the reduction of homicide rates.

So I decided to look by myself to the whole violence numbers in the state of Rio de Janeiro and see if the increase in killings by police officers reflected in an actual reduction of violent crime rate.

Data Analysis

I will focus the study in more violent crimes that need a closer look from the authorities: rape, kidnapping and civilian deaths in any criminal situation. The study will consist of an analysis of the time series trends, a correlation verification and time series forecasting. The variables are described below.

Legend:

• pol_kil is the number of people killed by police officers.
• rape is the official number of victims of rape
• kidnap is the total kidnapping cases in a specific period
• deaths is the number of civilians who lost their lives in criminal situations. This is a sum of all the columns of the dataset that involve civilian death.
• ano and mes are respectively the year and month in which the cases happened.

After the proper pre-processing the time series can be plotted in charts for an initial verification. Since data was not collected for all variables since 1991 some straight lines will be visible in some of the charts.

#time-series-forecasting #violence #data-science #brasil #time-series-analysis #data analysis