Skip to main content

Using data science to fight crime

Date

 

We expect our police forces to deal with an increasingly diverse number of problems – from tackling burglary and violent crime to safeguarding vulnerable communities and responding to critical incidents. While the nature of crime is changing, the tools and datasets we have for understanding and combatting it are also evolving.

Police agencies are data-rich organisations and data analytics is one tool that is becoming invaluable in supporting police to make the most of their information. The goal is to effectively and ethically deliver the resources they have in ways that maximise safety and minimise harm in our communities.

Researchers at LIDA are at the forefront of this new science, working with a number of different partners developing ways to use data analytics to better understand, predict and prevent crime.

LIDA is working with Safer Leeds, investigating ways to analyse free text data from crime reports – a process which requires natural language processing and machine learning techniques to look for patterns and similarities within the reports. These tools will help crime analysts better understand patterns in types of offending and aim to offer an early warning system for new emerging criminal behaviours.

“Vast amounts of rich unstructured text data are collected by police and their partners on a day-to-day basis,” says David Jackson, Partnership Intelligence Lead at Safer Leeds. “These large datasets present significant analytical challenges, but also offer huge opportunities. The work we’re doing with LIDA will help us harness this resource to better understand and ultimately, we hope, reduce crime.”

Dr Daniel Birks, Academic Fellow in Quantitative Policing and Crime Data Analytics at the School of Law and Fellow of the Alan Turing Institute, says: “A typical example of these new types of offending are recent spikes in thefts and muggings committed by offenders riding mopeds. Police intelligence has always had an operational understanding of these offences, but using data analytics we can proactively identify these new criminal behaviours more clearly and rapidly, and start to understand the types of conditions under which they’re most likely to take place.”

Working with the N8 Policing Research Partnership, LIDA has established an alliance across the N8 group of universities and 11 UK police services to address the challenges of modern policing.

Through the partnership, LIDA has teamed up with West Yorkshire Police to use machine learning to tackle vehicle crime. Researchers are developing ways of recognising vehicles with falsified number plates, by comparing the genuine make, model, year and colour of a vehicle with the records accessed via automatic number plate recognition systems. This will allow them to flag vehicles that are likely to have been cloned and might go on to be involved in criminal activity.

Taking a broader look at criminal behaviour, Dr Birks is also developing advanced computer simulations of crime using a technique called agent-based modelling. These models, which he describes as ‘synthetic societies’, can be used to better understand the link between the individual offender and victim behaviour and widespread trends in crime across society.

“Using these models, we can carry out experiments that would otherwise be impossible in the real world,” says Dr Birks. “We can use them to better understand how, for example, different street network configurations or neighbourhood facilities might increase or decrease people’s risk of victimisation.”

The team has recently received funding from The Alan Turing Institute to explore how these simulation techniques might also be used in practice. Working with partners at UCL, the Metropolitan Police and the National Police Co-ordination Centre, the goal is to see if tools can be developed to help police better understand the complex challenges of resourcing and demand.