A Study on the Construction of Estimate Models for Bicycle Theft Frequent Area

Abstract

In recent years, the crime cases known to the police in Japan are on the decrease, but public anxiety about crime remains high. In particular, the number of bicycle thefts cases is high, but the arrest rate is extremely low. Hence, it is a critical issue to take measures to prevent bicycle thefts. In this study, we constructed a model to estimate the incidence of bicycle thefts using data on the case of bicycle thefts for the past 7 years and clarified the elements of urban form associated with the occurrence of bicycle thefts. First, the optimal bandwidth for predicting density of bicycle thefts in the kernel density estimation method was verified, and it is shown that 40m is optimal to predict the density. The result of multiple regression analysis shows that numbers of free municipal bicycle parking spaces, numbers of passengers at stations, and areas of commercial district had a particularly strong relation to the density of bicycle thefts.