The Pattern of Crime Occurrence in South Korea Using Watson Analytics

Abstract

Crime has become a major concern in South Korea. OECD Better Life Index indicates that South Korea ranked 26th among 38 Organization for Economic Cooperation and Development (OECD) countries. This paper explores crime patterns in South Korea by using IBM's Watson Analytics. Identifying patterns can contribute to provide more efficient solutions to reducing crimes. Moreover, this information can be provided to citizens to help them avoid dangerous situations. About 180,000 data points from 2004 to 2017 were collected by the Supreme Prosecutors' Office of the Republic of Korea, Statistics Korea (KOSTAT), Korean Educational Statistic Service, and Korea Meteorological Administration. The paper focuses on the overall crime occurrences and felony crimes based on two major categories: situational information and offender's background information. Data were analysed using IBM's Watson Analytics, and several insights with criminal activity were found. The key findings are: 1) the higher the temperature, the higher number of sexual and violent crimes; 2) the percentage of male criminals has been consistently higher than the percentage of female criminal despite the proportions of male and female being roughly equal in South Korea; 3) crimes occurred more frequently in regions with higher population density; 4) the majority of criminals have no religion; 5) it is likely the offenders and victims do not know each other.