Page 8 - Case study: Crime prediction for more agile policing in cities – Rio de Janeiro, Brazil

2. The smart project(s)

Cities are where the future lies. They are hubs of innovation, productivity and experimentation.
However, cities also are sites of crime and violence. More than ever, municipal authorities, private
firms and civic groups are experimenting with new ways to improve safety in cities. In some cities,
new technologies are being deployed to improve the situational awareness of public authorities and
citizens. In others, all-encompassing surveillance and monitoring systems are being implemented,
raising challenges on the fundamental norms of privacy.

In most developed cities, high-frequency time series information on insecurity is increasingly
available. Literally thousands of gigabytes of raw data are available representing the dynamics
and characteristics of crime. New high-power computer analysis is giving rise to a next generation
of smart, agile and evidence-informed policing strategies. Predictive platforms in particular can
enhance police operations, identifying priority targets for police intervention, and enabling more
effective allocation of police resources.

2.1. Vision and content

Predictive analytics are hardly new. Statistical and mathematical models have long been used to
predict where crime may occur. Predictions are based on a series of assumptions. For example,
the criminological literature predicts that violent crime and property crime are not only highly
concentrated in specific locations, but also tend to occur at predictable intervals.

Predictive policing tools are being rolled out by police departments across North America, Western
Europe and parts of Asia. Police departments typically use thermal maps indicating the locations
and times where the probability of crime is highest. Senior law enforcement officers can apply this
information to plan their routine operations and send officers and patrol cars to the right locations
at the right time.

This so-called "hotspots policing strategy" - merging data analytics with targeted policing – has
been around for over two decades. Scientific evaluation studies have shown that it is an effective
crime prevention strategy . While concentrating resources on crime hot-spots may contribute
5
to a modicum of crime displacement, it happens less than expected. Indeed, departments must
regularly update their data systems and operational strategies as crime itself undergoes structural
changes over time .
6
Predictive policing is also evolving. They are benefiting from advances in machine learning, coupled
with more affordable computational power. When compared to traditional hot-spots mapping
approaches using retrospective data, predictive analytics can process more granular data at a more
rapid pace, generating predictions associated not only to a location, but also to a crime type, and to
specific times of the day and days of the week. When applied with fidelity, such tools can help police
departments validating their predictions on a daily basis and adapt their responses accordingly to
the everchanging crime patterns.

2 Crime prediction for more agile policing in cities – Rio de Janeiro, Brazil