The predictions are based on a detailed analysis of actual crime figures covering offences such as muggings, burglary and theft. Trials in Amsterdam have been so successful that the system is now to go nationwide.

The Criminality Anticipation System works by dividing the area to be studied into blocks of 125 metres by 125 metres. Each block is then analysed according to its crime history, the location of known criminals and other statistics such as the ages and sex of its residents, as well as how many people live on social security benefits.

Ethnicity

Ethnicity is not included as a factor because it is a ‘politically incorrect variable’, Dick Willems, who developed up the system, told the Volkskrant.

The results are then used by the police to carry out extra patrols or other preventative measures.

The trial in Amsterdam showed that 20% of burglaries took place within the highlighted areas, and this figure rose to 40% when the border areas were also included, the paper said.

Crimes such as murder and rape are not included because they are too rare to allow predictions.

Risks

Arnout de Vries of scientific research institute TNO said last year that big data has a role in policing but that mathematical models do have inherent risks.

‘The software performs calculations based on the collective memory of the police force, which is far more than any individual is capable of’ he said. ‘However, mathematical models also have inherent risks. Tunnel vision is one such risk, because the software relies on data collected in the past.’