DP半岛街头抢劫犯罪案件热点时空模式

The spatio-temporal patterns of street robbery in DP peninsula

选取H市中心城区DP半岛作为研究区域,以岛上2006-2011年发生的街头抢劫案件(共373起)作为研究对象,将DP半岛内街头抢劫案件的时空分布特征分别从宏观和局部微观两个尺度层面进行系统的分析。首先,对岛上的街头抢劫案件按年、月和小时进行统计分析,总结其在不同时间尺度上的变化规律:2007年开始的严打使案件数量逐年减少,直到2010年才略有回升;春节期间(二月前后)的案件数量明显高于其他月份;晚上22:00-23:00期间是案件高发时段。其次,利用Kernel密度方法对研究区街头抢劫犯罪的宏观空间分布进行整体的辨别,剥离出犯罪热点空间分布,分析热点与道路网和土地利用的关联性,结果表明热点多分布于主干道、通达性高的节点或土地利用混合度高的地方。最后,选出4个最主要的热点从微观尺度进行分析,PAI指数表明这4个热点在时间上是稳定的,从2006年到2011年一直存在。依据"热点时空类型矩阵"的时间分布和空间分布模式,将这4个稳定热点归类到不同微观时空模式,并对每类模式下的街头抢劫犯罪提出有针对性的防控对策,以便优化警力资源的配置、最大限度抑制和减少犯罪的发生。

The paper aims to uncover macro and micro spatial-temporal patterns of street robberies on DP Island of H city in southern China. DP Island, with an area of about 4 square kilometers, connects to the city via 7 bridges. There are a total of 373 street robbery incidences during the period of 2006-2011. Since street numbers are not linearly calibrated along streets in China, each incidence has to be geo-coded manually to a street network. First, all street robbery incidences are summarized by year （6 years）, month （12 months） and hour （24 hours）. The summary shows that street robberies reached a peak in 2007, followed by a trend of steady decrease until 2010; Just before the Spring Festival （arou^Ld February）, the number of incidences is much higher than that of other months. This pattern is different from those found in other countries, including the United States of America; during a day, street robberies peak during 22：00-23：00. Second, crime hotspots are revealed by kernel density mapping of all street robbery incidences. Comparisons with street network and land use suggest that hotspots tend to associate with main throughputs, intersections with high accessibilities, and areas with a high degree of land use mixture. To better characterize detailed patterns needed by policing, this paper selects 4 hotspots with the highest density for further analysis at a micro level. Prediction accuracy index （PAl） shows that these 4 spots are ＂hot＂ during all the 6 years from 2006 to 2011 and therefore warrant special attention for crime reduction and prevention. These hotspots are placed in different categories of the ＂hotspot matrix＂, which combines the spatial distribution （clustered or not） within the hotspot and the temporal distribution （clustered or not） in 24 hours. Based on these sp^Ltio-temporal patterns, the study suggests possible means for more effective policing, policing resource allocation, and crime prevention strategies. The study represents the first case study of its kind in China, and it sets the stage for future comparisons with other Chinese cites and foreign cities. Unfortunately, the name of the city cannot be revealed per confidential agreement on the crime data.