| Crime is a serious problem affecting people’s production and life,damaging the economic and social development.The crime data is a typical spatio-temporal data.Spatio-temporal visual analysis is based on spatio-temporal data to explore,analyze and demonstrate the spatial and temporal distribution.The crime pattern could be excavated from the big data of public security using spatio-temporal visual analysis technology,leading the users understand the distribution and changing trend of crime hot spots in the region quickly and intuitivel,probing clues,providing support for crime prevention and control.Based on the analysis of research status about the current situation of crime analysis software in china and abroad,and summary of the results in our team on cases spatio-temporal data warehouse,crime spatio-temporal data mining,crime pattern recognition and visualization of crime data,the paper uses multi-view collaborative visual analysis model to construct the spatio-temporal data visual analysis system for crime.The main work and characteristic are as follows:1)This paper studies the crime analysis,spatio-temporal analysis,visual analysis and visual analysis system to establishes the methodology of spatio-temporal data visual analysis system for crime,including the goal for crime pattern recognition,the main analysis method for spatio-temporal data mining,the show means for interactive visualization.2)According to requirements of crime visual analysis,a multi-view collaborative visual analysis model is proposed which contains a Visual Analysis Model(VAM)and a Interactive Collaborative Model(ICM).The corresponding process is divided into four steps which include data transformation,data analysis,visual mapping and visual drawing.Accordingly,the observer pattern is applied to realize the coordination of the related views in this interactive process.Furthermore,the architecture of multiple-view collaboration is designed based on"view-to-data" and "view-to-view" interaction Web patterns.3)The criminal space-time visual analysis system is carried out,realizing the management for cases data and basic spatial data.The cases spatio-temporal data is shown in three dimensionalities of space,time and attribute by the view of points map,stack-line chart,descarte heat map,spatial temporal cube,parallel coordinates,single dimension scatter chart with algorithms,such as ST-DBSCAN single dimension scatter,and so on.The relationships for different views are established by the index in their visual data to realize the dynamic linkage.4)The hot region,hot interval,hot place type and spatio-temporal hot spot for the cases of electrombile-stealing,pick-pocketing,burglaries,motorbike-stealing in Fuzhou City within 2 weeks are analysed by the system.Furthermore,the functions of multi-view collaborative and ST-DBSCAN are used to analyze the associated cases.These analyses testing the effectiveness of the system primarily. |
