As the old adage goes, hindsight is always 20/20. It’s only after a crime occurs that family members, friends, and law enforcement officials recognize the “signs” of a delinquent’s impending lawbreaking. Police departments and other law enforcement agencies are frequently reacting to crimes. Now, thanks to Big Data analysis, they can be proactive. They might even be able to stop a crime before it happens.
The analysis of enormous sets of data to reveal trends, patterns, and associations is the broad definition of Big Data. It helps corporations interpret consumers’ shopping habits, schools identify ideal potential students, and now it’s helping to reduce property crimes. For example, general Big Data investigations have revealed that crimes are likely to be committed around brand new retail stores that are about to open, and around buildings with signs posted that indicate that it’s going through renovations. This knowledge allows law enforcement officials to allocate their resources properly. In other words, to have more boots on the ground near those locations.
Big Data is able to prevent crime by identifying trends that human brains fail to recognize. The Durham police department in North Carolina is a prime example of effective data mining. Analysts discovered that 20% of the “shots fired” phone calls that came into the police station were coming from a mere 2% (roughly two square miles) of the city’s area. The police officers adjusted their standard operating procedures so that there were more officers in or close to that particular area. The result of this “hot-spot” analysis was a 39-50% reduction in violent crimes in four years.
Harvard Medical School also has a Big Data success story. In conjunction with the United States military, Harvard researchers analyzed nearly a million soldiers who served between 2004 and 2009. Based on 38 different data sources and 24 different indicative factors of future criminal activity, they deduced that 5% of the soldiers had a high risk of committing a crime. The prediction was significant. Although that group was only 5% of the 975,000 soldiers, they committed 36% of the crimes. Ideally future researchers will be able to use Big Data to identify high risk groups earlier and direct them to resources such as counseling, effectively preventing their crimes before they happen.
The police are using Big Data analysis and related tools like Hadoop Spark to interpret social media posts as well. Nowadays many agencies have a presence on sites such as Facebook and Twitter. Notification apps on mobile phones and wearable tech like the Apple Watch are also becoming more common. They use social media and text notifications to announce Amber Alerts when a child is missing, alert locals about nearby thefts, and even warn them about severe weather. But communicating over social media goes both ways. A department’s Facebook “friends” and Twitter “followers” use social media to report suspicious people and vehicles, crimes in progress, and they can post videos and pictures, too. Unfortunately, police departments whose budgets have been cut don’t have the necessary manpower to monitor and respond to all of those reports. That’s where software and skilled civilian analysts come in.
Like us flawed humans, Big Data will never be able to predict crime with 100% accuracy. It’s a tool, an exceptionally helpful resource, but it’s not psychic. However, it is responsible for drastic decreases of violent crimes in cities such as Durham, North Carolina. If it’s applied nationally with the same success rate, that would mean a potential of a 50% decrease in crime. You don’t need to be psychic to know that adds up to a lot of lives (and taxpayer dollars) saved.