ABSTRACT
Ratherthanaggregatingcrimedatatoareas,point-basedmethodswork withaseriesofpointsthatidentifyincidentlocations,thesiteswhere crimesoccur.Traditionallycrimeanalystswouldprepare'pinmaps'to examinegeographicalconcentrationsofcrimeincidents.Computerised point-basedmethodsbeginwithdigitalpinmapsandcomputestatistical measuresofspatialclustering.Oneofthemostwidelyusedpoint-based methodsforcrimeanalysisistheSpatialandTemporalAnalysisofCrime (STAC)system.STACworksbycountingthenumbersofcrimesthatoccur inoverlappingcirclesspreadevenlyacrossthestudyarea(Block1995). 'Hotcircles'arecirclesthatcontainthelargestnumbersofcrimeincidents and'hotellipses'identifytheareasofdensestcrimeactivity.Manylaw enforcementagenciescurrentlyuscSTACtosummarise,visualiseand analysethelargequantitiesofspatialcrimedatathattheyhandleonan ongoingbasis(Block1997).AlthoughSTACisoneofthemostwell-known point-basedmethods,analystshaverecentlyexaminedotherspatialclusteringmethods,includingtheGeographicalAnalysisMachine(Hirschfieldet al.1997),kernelsmoothing(McLaffertyetal.1999)andKnox'stestof space-timeclustering(Canter1997).TheCrimeStatsystem,developedby Levine(1999)includesasuiteofpoint-andarea-basedmethodsfor exploringcrimepatterns.
