Exploratory Regression Analysis of Crime Trends in Richmond, Virginia

Abstract

Crime negatively affects the safety and lives of citizens daily and has been one of societies’ worst byproducts. If one can identify the underlying influences that cause certain crime trends, then resolutions could be found to address these issues and to possibly curb future crime. This study observed the potential spatial relationships and distributions between crime locations, crime type, socioeconomics, and demographic for Richmond, Virginia. Richmond was divided into 148 neighborhoods and crime incidents occurring between January 1, 2020, and December 31, 2021, were analyzed. Crimes were separated into two dependent variables: violent crimes (assault, murder, sex offenses, robberies) and property crimes (larcenies and burglaries). These two variables were then analyzed using Exploratory and Ordinary Least Squares linear regression to determine if any of the socioeconomic and demographics statistics had a relationship with crime trends in Richmond, VA. The three explanatory variables that were found to explain crime occurrence best were population density, percent of the population renting housing instead of owning, and the percentage of the population with a high school level education or less. The Global Moran’s I test was used to determine if any of the two crime categories had any significant clustering. Only 34% of violent crimes and 18% of property crimes were supported by the three explanatory variables and both crime categories are considered to have random distributions. These results indicate that the socioeconomic and demographic variables used do not accurately explain crime trends in Richmond, VA.