竊盜為社會中常見犯罪類型,在世界各國更是所有刑案類型中犯案件數的最大宗。而在台灣,汽機車竊案更是所有刑案中犯罪率最高的類別。犯罪區域研究,如果能考慮空間異質性和空間自相關的特性,對於犯罪行為受到空間的影響,能較僅以行政區域劃分之量化統計數據,呈現更好的觀察與詮釋。本研究探討桃園市機車竊盜的犯罪地圖製作,以地理資訊系統為基礎,分別以宏觀層次與微觀層次兩個部分分析。前者以不同的統計尺度的面量圖繪製,探索犯罪地點的空間分布;後者先用核密度推估法產製機車竊案的熱點分布圖,並以全域型空間自相關,探討機車竊案地點是否有聚集的現象,再利用區域型空間自相關分析,推算出機車竊案的空間熱點,最後定義桃園市機車竊盜的長期熱區與近期的潛在熱區。
Theft is a common crime. The occurrence of theft has the highest percentage among crimes world-wide. Among all types of theft, motor vehicle theft has the highest rate in Taiwan. Crime analysis with spatial heterogeneity and spatial autocorrelation could provide better spatial reasoning of the crimes, than simply deriving quantitative statistics with administrative units. This study investigates the theft of motorbikes in Taoyuan metropolitan area through mapping the locations of reported theft. Based on the Geographic Information System, both macro and micro approaches were conducted. The former was carried out with a choropleth map to delineate the spatial distribution of the general trend. The latter applied kernel density function to generate maps of criminal hotspots. Global spatial autocorrelation analysis is conducted to evaluate whether there are clustered spatial patterns. Then, local spatial autocorrelation is utilized to derive the hotspot and identify both long-term motorbike theft hotspots and recently developed potential hotspots.