Skip to main content

Data as Risk Predictor for Temporally-Concentrated Crime Series

Date

The following research project is currently being undertaken by PhD Student Emily Sheard:

Aim

To ascertain which types of data are most reliable to predict the risk of victimisation for crime series where there is little opportunity for preventative intervention following an initial event.

Background

Although some variables are indicative of heightened vulnerability e.g poor security, such insights are of limited use to crime practitioners if data cannot be accessed which is able to describe their spatio-temporal distributions.

As such, prior victimisation has received lots of attention as an indicator of future risk, particularly for burglary dwelling offences, since local offenders are usually well-placed to return to previously targeted locations. (1)

Problem

The spatio-temporal parameters of some crime types e.g. 'car key' burglary, appear to be determined primarily by the location of desirable goods as opposed to the daily mobility patterns of offenders, thus making them more difficult to anticipate.

Evidence

Research indicates that 'car key' burglars will, on average, travel further than 'conventional' burglars to commit offences (2) and the less deprived areas are more likely to be targeted (3).

This is likely to result in transient hot spots as offenders seek to maximise profits in the minimum amount of time possible given the inherant risks of returning to unfamiliar areas.

Approach

Crime data will be analysed to identify the main characteristics of the 'car key' modus operandi and the findings will then be used to inform the choice of predictor sets.

The latter will be derived from both 'traditional' (e.g census) and 'contemporaneous' (e.g consumer-related) sources, thus facilitating analysis at different spatio-temporal scales.

Impact

  • Increased understanding of the subject area
  • More efficient tasking of police resources
  • Improved prevention/detection rates
  • Development of transferable techniques

Future Work

This could include the integration of outputs with known crime data to create a prospective crime model i.e an ongoing, proactive process of spatio-temporal filtering.

Note

The author recognises that the 'car key' burglary MO is vulnerable to technological advancements so the research is being conducted primarily as proof of concept.

References

1 Johnson & Bowers, 2004, European Journal of Criminology. 1(2): 237-255
2 Carden. 2012. http://www.crim.cam.ac.uk/alumni/theses/Carden,%20R.PDF
3 Shaw et al. 2010. International Journal of Police Science & Management. 12(3): 450-459