A crime data analysis framework with geographical information support for intelligence led policing

The manual crime recording and investigation systems in police stations all around the world are generating piles of crime documents which make storage and retrieval of reliable crime information extremely difficult as well as inefficient. Furthermore, investigators of central authorities have to manually search through these documents and communicate between the relevant police stations to obtain required information. Eventually, delays in information flow between investigators of crime incidents cannot be avoided. Sri Lanka Police too have been facing the same set of issues over many years. To get rid of pilling of large number of documents annually in police stations, Sri Lanka Police is allowed to destroy the documents related to solved crimes which are older than five years. This may destroy not only "closed files", but also very valuable information that can be used in future crime investigations.To overcome this problem, this paper proposes a web-based framework with geographical information support which contains a centralized database for crime data storage and retrieval. Geographical capabilities of the framework support not only spatial analysis but also provide an efficient solution to current manual crime map generation. Our highly secured and user friendly framework follows the state of the art layered architecture which provides an optimized data model for fast access and easy analysis of crime data. The solution consists of an affluent set of data mining tools which are essential in any crime investigation process. Security of data is ensured with data encryption for sensitive information and by limiting access to the data through a role based access method. Further the data is only accessible through a virtual private network (VPN) connecting all the police stations and other relevant departments of the Police. The proposed framework was evaluated by conducting an experimental study and the results are promising.