Spatio-temporal Features Identification And Visual Analysis Of Crime Cases

At present,our country is in the critical period of social transition and social contradictions prominent period.Urban crime,as a kind of social phenomenon is harmful to public safety,people's lives and property,having a stable upward trend,to grasp the space-time features of crime has great practical significance to prevent and combat crime,and to ensure the safety of people's lives and property.It is necessary to study on spatio-temporal variations of crime,which is significance to prevent and combat crime and to protect the people's lives and property.According to the lack of effective utilization of spatio-temporal information in the previous study of crime analysis and the lack of interactive expression in criminal information visualization,the study of data driven and business oriented to spatio-temporal analysis and visual expression of crime are carried out.Based on efficient spatio-temporal data mining method,the spatio-temporal distribution and patter of crime arc obtained.Multi-scale interactive visualization of crime spatio-temporal agglomeration is designed.The research results are as follows:(1)The study of crime spatiotemporal distribution based on pick-pocketing,electric vehicle theft and burglaries data of Fuzhou City in 2014 shows that these three types of cases overall present northwest-southeast diffusion trend and hotspots gather in the eastern area of Gulou.In space,these three types of crime are generally concentrated in the commercial area.In time,different types of crme have different seasonal trend and the distributions of different crime within 24 hours in a day also have obvious differences.In temporal and spatial distribution,the spatio-temporal hotspots of pick-pocketing are mainly concentrated in Taijiang walking street and Wan-Bao business district at noon and night.The spatio-temporal hotspots of electric vehicle theft are mainly concentrated in Wan-Bao business district before dawn.The spatio-temporal hotspots of burglaries are mainly concentrated in Information Vocational College and the surrounding residential district at 0 am to 2 am.(2)In view of the problem that criminal association rules mining ignore different data items with different importance and use MSapriori algorithm to calculate support of items with low efficiency.Multiple minimum supports combined with data cube model are adopted to optimize the issue above and multidimensional data association model is constructed for case event analysis.This model is applied to mine the crime pattern surrounding four three class hospitals in Fuzhou city,and then the results are showed in the form of thematic map and visualization.(3)Based on the depth study of crime spatio-temporal agglomeration and according to the requirements of the security services,this paper focus on interactive visualization and crime spatio-temporal agglomeration of interactive visualization is designed,which is including date scale,hour scale,urban district and crime scene.