本研究以大安區住宅竊盜犯罪案件為研究範圍，透過地理資訊系統製圖與集合性空間資料分析方法，針對「派出所轄區」、「里區域」、「自行制訂虛擬網格區域」等不同空間層次中犯罪空間分布狀況與集群態樣差之異加以分析比較，以作為警方犯罪熱區分析之參考。研究結果顯示，大安區住宅竊盜犯罪之空間分布與集群態樣於不同區域層次上存在相當之差異；差異區域層次犯罪空間分析方式可協助警方更確切掌握犯罪熱區空間分布，對犯罪預防實務工作之應用而言，「自訂虛擬網格區域」分析優於其他層次之分析，並更加貼近真實之犯罪熱區空間分佈態樣。

The study used data of residential burglary in Daan District from Jan. 2007 to Jun. 2009, and tried to explore issues of spatial distribution and clustering of burglary hotspots in different spatial levels of "village area", "police station district", and "virtual area" by using GIS crime mapping and aggregate data spatial analysis as major methods, in order to propose and simplify efficiency crime prevention policies for police organizations with their capacity. The results showed that aggregate data spatial analysis of different parameters setting can produce thematic maps which display spatial clustering of crime hotspots/ coldspots significantly. We then discovered that the residential burglaries were not randomly distributed, but significantly concentrated in certain areas of different area levels of Da-an District. Further, analysis of different spatial levels helps the police mastering spatial distribution of crime hotspots more and correctly, and the spatial clustering analysis in "virtual area" level is more close to real spatial distribution of crime hotspots than in "police station district" and "village area" level for practice of crime prevention.