Data Science and Death: The Race to Prevent Mass Shootings

How can the U.S. prevent more mass shootings?

That question has been asked countless times in recent years, and is being asked again after the terrible and historic terrorist attack at an Orlando nightclub that left at least 49 people dead on June 12, 2016.

The proposals about how to prevent such shootings have ranged from the ridiculously demagogic to thoughtful and detailed. The more reasonable responses have tended to fall into two categories:

  1. Improved detection of potential threats from mass shooters, including terrorists
  2. Stronger regulations and controls of guns

Predictive analytics and forecasting tools could be leveraged in either type of strategy, or in some mixture of the two. Let’s start with anticipation and prevention of terrorism. In recent years, there have been various initiatives for using data science and machine learning to stem such activities, including:

  • Aiding the CIA in discovering patterns and anticipating events in its recently established Directorate for Digital Innovation
  • Detecting suspicious behavior in live video feeds
  • Identifying terrorists via the V signs they make in videos
  • Predicting who might be a terrorist via cellular network patterns (which is problematic for a number of reasons)
  • Identifying potential cyberattacks
  • Creating an index based on web activity that might point to future terrorist attacks

In the case the Orlando attack, the shooter had previously been investigated by the FBI but was found not to represent a threat. He also legally purchased a semi-automatic assault rifle not long before the shootings. If both data points, as well as others uncovered by the prior investigations or from other sources, were known and used as input in a predictive analytics model, could a caution flag have been raised for that particular person? Even if that’s statistically and logistically possible, would society want such personal information collated and churning in a government algorithm somewhere? And do government agencies have the resources to follow up on such signals? These types of questions will increasingly be raised in coming months.

There are, in fact, groups and initiatives already trying to anticipate and stop mass shooters in the U.S. The Association of Threat Assessment Professionals recently held an annual conference at which there were sessions such as “20 Years of Workplace Shootings” and “Evil Thoughts, Wicked Deeds.” Journalist Mark Follman reports:

[T]hreat assessment teams can now be found everywhere from school districts and college campuses to corporate headquarters and theme parks. Behind the scenes, the federal government has ramped up its threat assessment efforts: Behavioral Analysis Unit 2, a little-known FBI team based in Quantico, Virginia, now marshals more than a dozen specialists in security and psychology from across five federal agencies to assist local authorities who seek help in heading off would-be killers. Those calls have been flooding in: Since 2012, the FBI unit has taken on more than 400 cases.

The goal of some initiatives is not just to anticipate mass shooters but to intervene through mental health counseling and other strategies so that a person who has the characteristics of a potential shooter does not, ultimately, become one.

One macabre side-effect of the sheer numbers of mass shootings in the U.S. is that the data on them is growing. This allows experts of all types to analyze the details of these events. The hope is that over time, we will get better at successfully predicting and intervening to stop such mass murders.

The other part the ongoing debate over mass shooting dwells on the subject of gun control, which is notoriously contentious in the U.S. Some experts point to the disproportionate emphasis on battling terrorism as opposed to reducing gun violence. John Hogan, director of the Center for Science Writings at the Stevens Institute of Technology, writes:

[B]etween 1970 and 2007, a total of 3,292 people in the U.S. were killed by terrorists, resulting in an annual risk of one in 3.5 million. Almost all those deaths occurred on a single day, 9/11/01. In contrast, more than 32,000 Americans are shot to death every year, according to the Brady Campaign to Prevent Gun Violence. Of those, more than 11,000 people are murdered and almost 20,000 kill themselves.

Hogan argues that if the U.S. devoted as much effort to gun control as it does to counterterrorism, the world would become a “much safer place.”

Others beg to differ. James Jacobs, director of Center for Research in Crime and Justice at New York University School of Law, argues that “there is no simple, effective policy to reduce gun crime that is just there for the asking,” especially in light of the U.S. Constitution’s Second Amendment and the Supreme Court’s decision that it guarantees the right of Americans to keep and bear arms. As for registration and background checks, he states:

There’s no universal registry of firearms, so if the police were to arrest somebody and try to prosecute whoever sold them their gun without the required check, there’s no way to verify who the seller was or when the sale took place. To have an effective system of regulating private sales you would need a registry, and the idea of a registry is an anathema to the gun owning community because they see a registration system as a precursor to a general confiscation—which it was in the U.K. and has been in other countries as well.

Even given a small taste of such debates, we can see why the issue of gun-related homicides in general and mass shootings in particular have remained such an intractable problem in the U.S. Predictive analytics, forecasts and threat assessments might be able to help prevent such shootings in the future, but they can’t be anything more than a piece of the puzzle unless political leaders find solutions that somehow balance individual rights with society’s needs to stem the tide of blood.

Data on the Larger Recent Mass Shootings in the U.S.

shooting data_1_edited-1

Subscribe to our mailing list

* indicates required


Leave a Reply