It’s name is KeyCrime and it is software used by the Milan police headquarters which has made it possible to reduce robberies by 57% and to discover the perpetrators in three out of four cases over eight years of field trials. Surprising results that have not gone unnoticed even abroad.
In fact, Peter Orszag, former collaborator of Bill Clinton and Barack Obama in the White House, has written about KeyCrime. Referring to the poor results obtained by the US police in solving the crimes, Orszag cited the Milanese software: “Reducing the discovery time of crimes is possible – he explained – in Milan, Italy, the police have been using KeyCrime for almost a decade. The software predicts robberies based on the day, time and place of the events that have already occurred, matching details of the criminals involved and their weapons”.
The results of the application for the solution and prevention of robberies were also studied by a researcher at the University of Essex, England, Giovanni Mastrobuoni, who calculated 2.5mln€ of damage has been avoided based on the fewer crimes that occurred and for the funds recovered thanks to KeyCrime.
There isn’t a group of researchers and super computer programming experts behind the invention of this application. It was created by a policeman from the Milan Questura, Mario Venturi, who independently began to think and create the software. Seeing its potential, the Milan police headquarters supported Venturi’s ‘research’ and, since 2009, has been using the application. The program can be used for all serial crimes, but for now it has only been applied to robberies.
But how does it work?
The policemen collect, as they have always done, the reports of the robberies which are passed on to the office in charge of data collection and analysis.
Here all the details are entered into the KeyCrime system and often the agents call the victims back to ask for further details, for example in which hand the bandit was holding the weapon or descriptions of disguises and build. At this point the software enters into action. Through algorithms, it finds correlations between the various criminal events and checks whether there is a possible possible connection – a crime series in other words. If the response is positive, a nickname is given to the as yet unidentified bandit believed to be responsible for the various shots.
But KeyCrime doesn’t stop there. In fact, it has the ability to predict robberies: on the basis of the data entered, it calculates the most at-risk targets in a given area of the city, the day and time when the bandit is most likely to take action. In this way, a targeted operation set up involving plainclothes policemen. The result is that the robbers are very often arrested in the act or the presence of the police puts them on the run before they go into action.
The results obtained by Key Crime are remarkably significant. In 2008, the year before the software began to be used, there were 664 robberies against pharmacies, shops and supermarkets in Milan, in 75% of cases the perpetrator was not caught.
In eight years of using KeyCrime, the numbers have reversed: now 74% of robberies have a known culprit and in 2015 the number of crimes scored dropped to 283.