Predicting Future Crime With Big Data
Updated: Jun 17, 2019
Historically criminals were hauled off to the gallows for what today would be considered minor offenses. The rope was slipped around the neck of the convicted pickpocket as well as the convicted killer. Both fell through the same trapdoor. The executioner worked his art without discrimination.
The New York Times correspondent Sandra Blakeslee reminded us that in 1765, John Ward was hanged for stealing a watch and a hat.
The historical cases reveal a very different world of criminals and law enforcement officials. The authorities have been reactive. They’ve had to wait until a crime has been reported before springing into action. Catching someone who violated the law meant rounding up witnesses and gathering evidence that implicated a wrongdoer.
The old policing model has very little to say about the future. It functioned on what was known in the present. A victim lodged a report. It also rested on the hunch or intuition of the police. Experienced police had knowledge about neighborhoods. Though that information, in the large scheme of things, was bound to be incomplete and tainted by bias. Until recently, the literature of crime followed the Sherlock Holmes model of a logical, clever and objective detective who outsmarted the villain.
We inhabit a very different world now. Not only do most countries no longer hang watch and hat stealers, they are using Big Data to predict geographical areas where crime may, on probability, be more likely to occur and with that information police can step up patrols. We have entered the machine age of law enforcement. The old model is in the process of a radical change as Big Data arms the police with predictive models and that takes policing into the future where crime hasn’t yet been committed. Such a change allows for development of policies of crime suppression for crimes that might be committed.
Los Angeles police have managed to reduce burglaries (33%), violent crimes (21%) and property crimes (12%) by adapting software developed to predict earthquakes and aftershocks. Eighty years of crime that included 13 million criminal acts were fed into the mathematical model that used the data to predict the areas where crime was most likely to occur. It seems the model yielded good results. New crimes are constantly added to the database, and the LAPD officers who were at first resistant to taking orders from a mathematical model have become true believers.
Chicago police have gone beyond hot spots to using Big Data to target people most likely to commit a crime in the future. There is mapping of crime hot spot areas of the city and the mapping of social networks is the logical extension. “Commander Steven Caluris, who also works on the CPD’s predictive policing program, put it a different way. ‘If you end up on that list, there’s a reason you’re there.’” In the future, the map of your social network may be used by the law enforcement agencies to assign you a probability statistic as your future criminal activity. Like a travel ban list for air travel, you may never know what is behind the inclusion of your name on a hot list. Florida is going down the same road.
Professors at Rutgers’ School of Criminal Justice have received grants to develop software called Risk Terrain Modeling Diagnostics Utility. Here’s a glimpse of the future of Big Data in policing:
“The National Institute of Justice recently awarded two grants, totaling nearly $1 million, to conduct RTM research in seven U.S. cities: Newark; New York City; Chicago; Arlington, Texas; Colorado Springs, Colo.; Glendale, Ariz.; and Kansas City, Mo. Researchers from Rutgers’ School of Criminal Justice and John Jay College of Criminal Justice at the City University of New York are conducting the studies using the RTMDx Utility. The Rutgers software is currently being used in the top four U.S. markets: New York, Los Angeles, Chicago and Miami. It is being adopted by industry and law enforcement offices in many countries, such as Australia and Canada, and major foreign cities such as Paris and Milan.”
The Australian Crime Commission has also funded a big data project. The goals is to use to “data mining to trawl through data sets looking for patterns and potentially predicting emerging crime issues and trends across the country.”
The promise is that patterns emerging from the big data will allow the police to identify areas where resources are needed. This has the advantage of consolidating resources in the areas where crime is most likely. It is being sold on the basis of efficiency. Like Wall Street brokers, the police have entered the world of big data with the goal of assessing risks. For a broker, it is getting in and out of a stock so as to make a profit. For the police they have structured data that predicts what types of crimes are on the increase or decrease for a given geographical area. The police study the big data looking for trends. And like a broker, the police having identified a trend, can allocate necessary resources to deal with the kinds of crime that are predicted from the data.
The BBC reported on Big Data in crime prevention, noting the need to accumulate masses of data about an area in order to predict crime trends. Making connections between crime and connections, and those that happen across international boundaries leads to unraveling complex networks of individuals. The BBC report shows how far we’ve come since the hanging of John Ward in 1765. Big Data allows a corporation to detect who on the inside is communicating with whom on the outside and to look for patterns that suggests an employee may be leaking information. It also allows the military tactical advantage in the field as Big Data is constantly fed into analytical models updating positions, movements, and communications on the ground.
Philip K. Dick predicted in The Minority Report that the State will evolve a system to predict crimes before they are carried out. The Big Data is used to define ‘hot spots’ where crime is most likely to occur. In the future, before you buy that house or condo, you might want to ask the real estate agent about whether the property is within a crime hot spot!
One should bear in mind that we are very early days into collecting and mining Big Data. The dynamics of technological change make predictions in the medium and far future nearly impossible. The reality is that we are headed down a road for future decision-making about the mechanism of the criminal justice system and we don’t know where it will lead us. We only have best guesses and cognitive biases such as best-case scenario. We run the real risk of an information infrastructure that will build a criminal justice system that surrenders our notions of free will and liberty.
In the future, John Ward may be hanged before he steals the watch and hat, doomed by Big Data, which assigns a 98% probability of future criminal conduct. Or if he had a 98% probability of being a serial killer, would you agree that he should be arrested and sent to prison? On the Big Data road map, this might be a destination. We have set out on a long journey and along the way we lose much of what we value as individuals for a class of elites who have most to gain in a new culture based on total security.