Chair: Wim Hardyns

WG-PLACE: Methodological innovations: emerging approaches to studying crime and place

Building: A
Room: 01


Author: Engström Alexander, Malmö University

Karl Kronkvist, Malmö University
Title: Stunda: Examining Experiences of Situational Fear of Crime Through Smartphone Applications Among Young Adults in Malmö
A situational dimension is often discussed in the fear of crime research but most studies rarely adopt a genuine situational data collection methodology. However, recent technological developments and increased smartphone usage have contributed to innovative methods and better opportunities to collect adequate situational data. Drawing upon the inventive approach by Solymosi, Bowers and Fujiyama (2015) this presentation will give insight into an ongoing research project where the feasibility of collecting information about experiences of fear of crime by using a smartphone application among young adults in Malmö (Sweden) is examined. The smartphone application, STUNDA, aims to collect traditional survey data but is also designed to gather situational data through Experience Sampling Method (ESM). ESM research encourages study participants to answer questions related to experiences and feelings as they occur in a specific point in time and the methodology is well-suited for integration in a smartphone application. The participants are requested to report their experiences of fear of crime at the specific moment as they receive a push notification but they can also report experiences retrospectively. By collecting data about experiences of fear of crime and provide these with a spatiotemporal stamp, this situational research may contribute with new knowledge of relevance for fear of crime research.
Keywords: situational fear of crime; smartphone; application; experience sampling method
Author: ,

Title: Using Real Data to Simulate Offenders Within a Virtual Environment
Keywords:
Author: Rummens Anneleen , Ghent University and The Institute of International Research on Criminal Policy (IRCP)

Wim Hardyns, Ghent University and The Institute of International Research on Criminal Policy (IRCP)
Title: Comparison of Near-Repeat, Machine Learning and Risk Terrain Modelling for Making Spatiotemporal Predictions of Crime
Decision-making processes are increasingly guided by intelligence gained from predictive analysis. In the context of crime data analysis, predictive analysis methods (commonly called predictive policing in this context) are used to make spatiotemporal predictions of criminal events. When looking at current applications of predictive policing, three main groups of methods can be distinguished: near-repeat modelling, machine learning modelling and risk terrain modelling. Using crime and socio-economic data from a large city in Belgium, a comparison is made between the prediction performance of a near-repeat model, a machine learning model and a risk terrain model. The results of this analysis and the implications of our findings are discussed.
Keywords: predictive policing; predictive analysis; near-repeat modelling; machine learning; risk terrain modelling
Author: Salman Haleem Muhammad, Manchester Metropolitan University Crime and Well-being Big Data Centre

Monsuru Adepeju, Manchester Metropolitan University Crime and Well-being Big Data Centre
Title: The Scale, Nature and Spatio-Temporal Patterning of Mental Health-Related Incidents: Insights From Text Mining Police Incident Logs
Recent years have witnessed a dramatic shift in the landscape of policing in the UK, with new forms of non-crime demand increasingly prevalent. The College of Policing (2016), for example, has estimated that between 20% and 40% of police time is spent dealing with mental health-related incidents. However, only a very small proportion of incidents are typically flagged as mental-health related, meaning that little of the true scale and nature of mental health-related demand, inclusive of its spatio-temporal dimensions, is known to police forces. This serves as a major barrier to developing effective and efficient place-based interventions. To overcome this shortfall, we deploy text-mining strategies to extract information from incident text logs, enabling more accurate quantification and, via Latent Dirichlet Allocation, qualification of mental-health related incidents. Thereafter we deploy Kernel Density Estimation to evaluate their spatio-temporal clustering. The paper concludes by considering the implications of this research for future police incident recording practices and place-based policing interventions.
Keywords: text mining; mental health; incident logs; demand
Marked presentations to PDF Generate ics file