Bayesian Spatial-temporal Modelling and Mapping for Crime Data in Nairobi County

Abstract

Nairobi is a county in Kenya that is more prone to crime occurrence. This has made many researchers, for the past years, to study about crime occurrence in its suburbs and which factors promote crime. The theories around crime are always coupled with an attempt to predict their occurrence, for better crime analysis, and management, in case they happen, the associated covariates and their changes are analyzed. At the sub-county level, the crime occurrence is highly studied and understood. In this study, using Bayesian theory, this study builds spatial-temporal Bayesian model approach to crime to analyze its spatial-temporal patterns and determine any developing trends using data regarding robberies that occurred in Nairobi County in Kenya from January 1, 2011 to December 31, 2018. Of the diverse socio-economic variables associated with crime rate, including unemployment rate, poverty, weak law enforcement, Alcohol and drug abuse, and illiteracy, this study finds that robbery crime rate is significantly correlated with the poverty index and the unemployment rate. This finding provides a statistical reference for County safety protection. For further work, we recommend that further study can be done to determine factors associated with the dynamics and the distribution of crime in Nairobi County while accounting for measurement error that might be present in the covariates.