Computational techniques for the geo-temporal analysis of crime and disorder data.

Abstract

This thesis presents research that has resulted in the development of techniques and methodologies through which crime and disorder can be analysed. These are applicable for use by law enforcement agencies and those charged with the management of crime, injury and community safety. A fundamental element in developing the techniques presented in this thesis was an appreciation of those currently in use. National surveys of all three emergency services (Police, Fire and Ambulance) offered an insight into the use and uptake of computerised mapping technologies for both operational and strategic purposes. The surveys highlighted current trends in and attitudes to such technologies in addition to the identification of potential future developments. Further insight was gained through working with local Crime and Disorder Reduction Partnerships. Here a framework was designed and implemented by which crime and disorder data could be audited using geo-statistical techniques.

Two key exploratory techniques are presented in the thesis that develop the currently available tool set. The first, uses animation embedded within a Geographical Information System through which multiple snap shots of criminal activity can be aggregated and played back to the user in the form of an animated sequence. The second, develops the use of animation to provide a way by which observed movements (or flows of criminal activity) can be qualified through analysis of clusters over time. This is achieved by quantifying volume, centroid and direction of movement.

The effective allocation of finite resources is a key issue for each of the public services covered by this thesis. A novel methodology is proposed to address this requirement. The methodology uses a geographical scanning algorithm to generate clusters of sufficient data from which Artificial Neural Networks can be trained to model a simple cause and effect relationship.

Date of Award	2003
Original language	English
Supervisor	David Kidner (Supervisor)