Crime continues to cast a shadow over citizen well-being in big cities today, while also imposing huge economic and social costs. Prevention, early detection and strategic mitigation are all critical to effective policy intervention, especially in domains where coordinated responses are required.
Every day, about 10,000 incidents are reported by citizens, recorded and geo-referenced in the London Metropolitan Police Service Computer Aided Dispatch (CAD) database. Today, impending funding cuts bring new pressures for central accountability and improved efficiency, while community empowerment initiatives bring new opportunities and challenges to policing. Timely understanding of how criminality emerges and how crime patterns evolve is crucial to anticipating crime, dealing with it when it occurs and developing public confidence in the police service.
It is widely understood that policing, crime and public trust all have strong spatial and temporal dimensions. An integrated approach to space-time analysis is needed in order to analyse crime patterns, police activity patterns and community characteristics, so as to understand and predict the when, where and what of how criminal activities emerge and are sustained.
This research will consolidate achievements in integrated spatio-temporal data mining and emergent network complexity to uncover patterning in crime, policing and citizen perceptions at a range of spatial and temporal scales. Each dataset of police movement, crime (and disorder) reported in the CAD, and citizens making ‘999’ calls constitutes a spatio-temporal network (STN), which has its own characteristic patterning and behaviour in space-time, and which interacts with the other STNs.
The (geotagged) deployment of police manpower in space and time, the spatio-temporal patterning of crime and disorder, and the perceptions of members of the public are likely to be interlinked to differing extents. The first of these purportedly both anticipates and responds to the second, while the third is a lagged response to the first two, giving reason to anticipate that all three networks should be tightly coupled.
The project will first analyse spatio-temporal patterns of individual STNs, then associate the patterns among these STNs via integrated spatio-temporal data mining developed using innovative statistical regression and machine learning.
This research will utilise a range of disciplines (crime, geography, geoinformatics, and computer science) to help engineer effective practical solutions to crime problems. It proposes a new method for exploring crime patterns and integrating information on crime and police activity. It systematically addresses a structured programme of analytical issues in spatio-temporal data mining, which are becoming core to Geographical Information Sciences. It will advance the theory, methodology and application of research into network complexity by evaluating the forms and interactions of the networks that characterise crime and other socio-economic phenomena. This will make it possible to not only understand activity networks but also to use them for prediction and decision making.
This addresses the aims of RCUK’s Global Uncertainties Programme in crime, terrorism, and ideologies and beliefs. It will extend our appreciation of the subtle interplay of different forms of complex systems, in ways that will contextualise tactical and strategic responses to terrorism and organised crime. It will enable intelligent policing of London Metropolitan Police by granting unforeseen levels of prediction. The best practice of individual Metropolitan boroughs can be extended to others in the UK.
The methodology developed here will be transferable to other international cities using similar incident report systems. This will directly benefit people who live, work and visit London and those cities to make them feel safe.
