Crime / Law

Data in Crime Fighting: Beyond Minority Report

27th Jun `14, 03:54 PM in Crime / Law

When we discuss Big Data in crime fighting, the analogy of Minority Report, the 2002 Tom Cruise film,…

Guest Contributor

When we discuss Big Data in crime fighting, the analogy of Minority Report, the 2002 Tom Cruise film, always comes up. This is the idea that it would be possible to predict who is going to commit a crime and when, meaning that law enforcement can stop these crimes before they are committed.

This is not a particularly accurate way of describing how Big Data is being used in society to help with crime prevention. Although there are ongoing experiments around the use of this kind of data to predict who could commit certain crimes (most notably the FAST programme currently being used by Homeland Security in the US) these kinds of initiatives only have a 70% accuracy rate, meaning that three in ten people who would be arrested, would have done nothing wrong. The idea behind this can also never be accurate, often crimes take place due to a rush of blood to the head, crimes of passion have very little premeditation meaning that unless government departments can read somebody’s thoughts, then these would be totally unpredictable.

Big Data has a much bigger role to play than a sci-fi version of policing, it has been making society a safer place (albeit in smaller ways) for the past 3 years.

It is well documented that US cities such as San Diego and New York have utilised the kind of data that new systems can create to not necessarily predict crimes in themselves, but pinpoint where crimes are likely to take place and preventing them. This could be anything from increasing a police presence in the area to changing certain physical aspects, improving street lighting or increasing the amount of visible CCTV in the area. This has been successful in both of these cities, allowing for police leaders to strategically place their forces in order to have the greatest societal good.

This kind of ‘predictive policing’ has also been adopted elsewhere in the world, with the UK and other European countries utilising similar systems to improve their own targeting of high crime areas. It has been a success and does not have the invasive elements of data collection that many would associate with the use of analytics in policing, it has been successful without people feeling like their privacy has been compromised.

This kind of information is relatively easy to collate however. Simply put together which crimes have historically happened in these areas and at what times. Therefore the likelihood of a certain crime happening in a certain area, at a certain time is x% higher than the average.

The difficulty with this system is the necessity firstly of crimes being reported and secondly the use of historical data, which for many areas may be lacking. This is why Rutgers have created Risk Terrain Modelling (RTM). RTM allows for crimes to be predicted not purely from the history of crime in that area, but from the surrounding environment and the likelihood of these conditions allowing for an increase in a certain amount of crime. There are always going to be the more obvious places, dark alleyways or areas that attract an increased amount of foot traffic, but this system allows for public officials to identify areas where crime is likely to happen that many would not consider.

The system has been made available to public bodies through an app that automates the processes without the need for additional crime analysts, meaning that police forces can make informed decisions about asset deployment.

This kind of work by Rutgers and others is incredibly useful for crime prevention and works well for demonstrating the power that data can have in deployment and keeping the public safe, but does not tell the full story of how a collaborative use of data can be invaluable in the actual solving of crimes.

A telling sign of how difficult it can be to track and solve crimes is simply the ways that criminals work. They often do not even use money in trading, exchanging hard to track goods as a substitute. How is it therefore possible to track the movement of something that isn’t known to exist? This becomes even more complicated when the ‘currency’ used is often switched across international borders, making collaboration difficult across police forces.

One of the most common forms of non-money-based currency is firearms. The use of firearms across borders is difficult to track as the actual guns traded are often untraceable through things like serial numbers and do not exist themselves on a database, at least not in their physical form. However, as firearms essentially have one use (i.e shooting somebody/something) it is possible to identify a gun and the links it has between countries through the distinctive residues and markings that it leaves. Through a new database (Odyssey) in the EU, countries can now track a gun through the shootings it has been involved in. The distinctive aspects of a particular gunshot can be noted in Italy and if those same characteristics appear in Britain, the chances are that this has been a cross border trade of the same gun.

This not only means that crimes can be easily identified and tracked but also gives law enforcement agencies the opportunity to pinpoint particular illegal trade routes. If there is an abnormal number of shootings with the same guns taking place in both Paris and Lisbon for example, the chances are that there is a trade link between criminal elements in France and Portugal.

A similar story comes from child protection elements of law enforcement.

When catching a paedophile it is common to find thousands of abusive images on their computers, sometimes amounting to hundreds of gigabytes of data that needs to be stored and analysed. The time it would have previously taken for a human to analyze all of the data, as well as making links between these images and others found in the same country let alone across borders, made the task beyond the single case almost impossible. Therefore systems are now in place that allow these images to be analysed and identified, not only throughout the crime in hand but also across other cases nationally and internationally.

Due to the nature of paedophiles and the networks that are created, this kind of identification of images and where they have been shared allows for law enforcement to identify rings and make multiple arrests rather than just one.

Big Data is not only used online for these kinds of actions either, they also hold power in the prevention of pirated software. At the Microsoft Digital Crimes Unit, they can identify when a serial number for a piece of software has been stolen and used. They can even pick out when it is being tested prior to it’s sale by the counterfeiters. This means that this kind of software can be pinpointed and appropriate action taken.

The use of data when finding the appropriate action in this case is also interesting. Through analysis it became apparent that it was far simpler and effective to follow the money to merchant accounts which allow online credit card transactions. Disabling these accounts meant that it didn’t matter how many websites were created to sell the counterfeit software (often multiple sites were linked to one merchant account), if the counterfeiter couldn’t take payment, they couldn’t sell fake software.

Similar to this is the use of data analytics in the detection of fraud in both insurance and finance, reducing the costs of premiums and allowing banks to reduce losses and pass on the savings to customers. A prime example comes from the UK ,where the Durham police force used analytics to identify a complex fraud that involved multiple claims from the same car crash. Through the use of data they identified who was responsible and how the kind of scheme that had been used had not only subjected the insurance companies to fraudulent claims but also increased the cost of insurance for the local population. Using a similar system to this, but through an internal team, Nationwide building society in the UK have managed to reduce the amount lost to fraud by 75%.

The use of Big Data in fighting, preventing and identifying crime has a huge knock-on effect to society in general. If criminal elements of society felt that they could lose themselves in the cloud of information or avoid identification through cross border work, Big Data has proven that this is no longer the case. The fact that this is a system that is being implemented so widely and has gained many fans across the world suggests that this is not simply a fad that police forces are jumping on for the sake of seeming progressive, but a genuine solution to many of the issues that have a detrimental effect on society as a whole. Although we may never hit the levels shown in Minority Report, the reality is that our current level of use is still making a big difference to how we conduct our police work.

This article was published in the 9th issue of Big Data Innovation Magazine. You can download the magazine from here.