Geographic profiling: how using police data on suspects’ activity locations might help solve crime

Ms Sophie Curtis-Ham1, Professor Wim Bernasco2, Dr Oleg Medvedev1, Professor Devon Polaschek1

1University Of Waikato, Hamilton, New Zealand
2Vrije Universiteit Amsterdam, Amsterdam, The Netherlands

Abstract:

Background: Research into offenders’ crime locations has highlighted that they are most likely to commit crime near places they conduct their routine activities. This research suggests that if we could map offenders’ routine activity locations then we could predict where they are most likely to commit crime. And if we could predict where individual offenders are most likely to commit crime, then we could predict who is most likely to have committed a crime, given its location.

Purpose: This study examined the predictive accuracy of a novel ‘geographic profiling’ (GP) process that operationalises these predictions using data readily available in police investigations.

Methods: The GP process identifies and ranks suspects based on their relative likelihood of having committed an input crime given information about their recorded activity locations. To test the accuracy of this process, we ran it for random samples of 1000 solved crimes, for each of five typically ‘whodunit’ crime types (residential burglary, non-residential burglary, commercial robbery, personal robbery and extra-familial sex offences). We then calculated how frequently the GP process accurately places the actual offender among the top ranked suspects, from a total dataset of >25,000 potential suspects.

Outcome: The GP process vastly outperforms chance, but more importantly is accurate sufficiently often to have promise for crime investigations in practice. We also identify factors that affect accuracy, such as the type of crime.

Conclusion: Operationalising the processes underlying offenders’ crime location choices through this geographic profiling method has potential to help police solve crime.


Biography:

Sophie Curtis-Ham, MSc, LLB(Hons), is a PhD Candidate at the University of Waikato and senior researcher at New Zealand Police. She has previously worked as an intelligence analyst in the UK and NZ. Her research interests include the development and application of the New Zealand Crime Harm Index, evaluations of policing initiatives, and the use of statistical methods for behavioural and geographical profiling to support criminal investigations.