Evaluating the effectiveness of big data policing: a randomised controlled trial in three Belgian police departments

01 October 2021 → 30 September 2025
Regional and community funding: Special Research Fund
Research disciplines
  • Social sciences
    • Causes and prevention of crime
    • Criminography and methods of criminological investigation
    • Police administration, procedures and practice
    • Safety, prevention and police
RCT crime Big Data machine learning police
Project description

‘Big data policing’ is an innovative intelligence-led policing strategy making use of historical data to forecast when and where there is a high risk of new crime events, with the objective to use police resources more efficiently and proactively, and ultimately reduce crime rates. Big data policing applications commonly analyse data using machine learning algorithms, with the aim to “learn” from past crime patterns to be able to forecast future trends and patterns. Big data policing has been introduced in several police departments around the world, with many more interested in applying it in the future. Despite this popularity, today only limited scientific evidence is available for the effectiveness of big data policing. So far, only three randomised controlled trials (RCTs) have been conducted and published, with mixed results, and all of them took place in US (metropolitan) cities. Therefore, there is a need for RCTs, with a specific focus on (1) European police contexts, (2) mid-term effects (≥ 1 year) of big data policing and (3) secondary effects such as user-experience of police officers, cost-effectiveness and ethical responsibility. The current project wishes to contribute to closing these gaps by conducting an RCT in three police departments in Belgium.