地理資訊系統應用於毒品查獲空間分布：縣市毒品查獲地點的分析

Application of geographic information systems to the spatial distribution of illicit drug use: an analysis of the location of illicit drug seizures in a city

目標：本研究針對三、四級毒品查獲地點，以地理資訊系統探討毒品查獲與地域環境的關係。方法：以某縣巿A從2011至2015五年間毒品查獲案件地點，使用ArcGIS進行空間分析。先進行全域分析，再將查獲率高低以視覺化圖形呈現。然後進行區域分析，包括二種熱點分析（Getis-Ord Gi*與核密度函數）。最後畫出毒品查獲數之年變動率圖。結果：5年間的平均最近鄰指數R皆遠小於1，顯示毒品查獲地點有空間聚集性。區域分布圖從只有少數高查獲率區，逐步蔓延成多區皆屬高查獲率區。南邊行政區域之查獲率屬中高等級者，其Getis-Ord Gi*的檢定也呈顯著而可判為熱區。套疊實際街道圖，用GeoPDF的格式輸出，可研判查獲地點與街道特性之相關性。核密度分布圖也顯示核密度值較高的點皆位於西南邊與鄰近幾區內。查獲數之年變動率與查獲率分年曲線圖，呈現各區毒品查獲之變動情勢不同。結論：透過地理資訊系統的視覺化歸納，配合熱點分析及風險圖像，有助於辨識出三、四級毒品的聚集點特徵，進而落實毒品防制的在地化策略。

Objectives: To investigate the relationship between the use of Category Three or Four narcotics (illicit drugs) and geographical characteristics of the locations of seizures using the Geography Information System (GIS). Methods: Using the seizure data from a city over five years (2011-2015), we conducted a series of spatial analyses using ArcGIS software. We started with global analysis, and determined the seizure rate in different districts using visualized figures. Local analysis was then performed using two kinds of hot-spot analyses: Getis-Ord Gi* and kernel density analysis. Finally, the change rates in the number of seizures over the years for individual district were determined. Results: Estimates of the average number of nearest neighbors showed that the ratio Rs were all less than 1 for the five year period, indicating a strong spatial aggregation of the seizure of illicit drugs. The district-specific seizure rate showed that ＂high-rate areas＂ started with a few and then spread to others. Local analyses confirmed that some southwest districts could be designated as hot spots since their Getis-Ord Gi* reached statistical significance. Combining the seizure locations with street maps in Geospatial PDF helped identify street characteristics associated with the use of illicit drugs. The results of kernel density analysis also indicated that the spots with the highest densities were located within the southwest districts and their neighbors. In addition, the rates of change in the number of seizures by year differed in individual districts. Conclusions: GIS-based visual displays with hot-spot analysis provided clues to the identification of geographic characteristics associated with the seizure of illicit drugs. These may enhance location-tailored strategies to prevent the use of illicit drugs.