Design And Implementation Of Crime Predictive Analysis System Based On Spark

With the rapid development of the internet of things,cloud computing,and mobile internet,big data has become the dominant direction for the development of information technology.At present,police department is faced with the problems of rapid data growth,various sources of collection,high processing delays,weakness analysis capabilities,early warning inefficiency,and analytical results ambiguously,making it difficult to meet the business requirements.This paper takes crime prediction as the research goal,adopts big data technology,designs and implements an early warning analysis system that can meet the needs of police department?s daily work and provide scalability for future development.The system has great guiding significance for the crime investigation under the background of big data,it improved the working efficiency of police department,and meet the development requirements of police department.Based on the deeply research of big data technology and analysis of the needs of police department,this paper designs a crime predictive analysis system based on Spark.With the rich data of police department,Firstly,it analysis the current status of police department,and confirm the functional requirements and non-functional requirements.Secondly,it develops the system design and implementation.It designs a variety of interfaces to collect multi-source data,and using message queue to aggregate data resources,and rely on Spark real-time processing components to refine the crime prediction data,and then the processed data will be judged by a machine learning model trained by historical data to get crime occurrence prediction result,the result will be visualized in GUI.And finally we carry on system testing.This system adopts a loosely coupled,extensible,and scalable distributed architecture,establishes a real-time and efficient data analysis system,and realizes real-time acquisition,real-time processing,real-time analysis,and real-time visualization of big data resources.The system is applied to the real situation of crime prediction,and established the detection mode of the person-to-case,make the working efficiently.