How Data Filtering Can Prevent Terrorist Attacks

The city of Nice has been called the video surveillance capital of France. The city has 1256 cameras that can be viewed from a central command center staffed by 70 workers. Yet the cameras did not stop Mohamed Lahouaiej-Bouhlel from driving a truck into the crowd of people gathered on the Promenade des Anglais to watch fireworks for Bastille Day, killing 84 and wounding 303 on July 14, 2016.

“The people who are monitoring video surveillance are already dealing with data overload. Just because you can generate more data doesn’t actually give you a solution,” says James Chong, founder and CEO of Vidsys. A real solution requires the right blend of people, process, and technology, Chong says. His company is helping monitor the Republican National Convention in Cleveland.

Algorithms, however, excel at collecting and analyzing and analyzing big data sets based on user-programmable rules. In Vidsys’ case, the company analyzes video data based on time, location, duration, frequency, and type. Its software can also weave in relevant data from other sources, including radar, building management, social media, and so forth.

In the example of the Nice attack, the time the truck began driving towards the boardwalk would have been one red flag. This event happened close to 11 p.m. during a public holiday in France. “In terms of filtering, we could set up a rule that says: 'After 9 p.m., we want to monitor any large cars or vehicles—whether it is an 18-wheeler truck or a delivery vehicle,'” Chong says. “And each time one is spotted, a situation is created.”

Location is important, too, especially for events such as the Bastille Day gathering in Nice or the Republican National Convention in Cleveland, which gather large crowds. “You can set up electronic boundaries—a perimeter around events,” Chong says. “In the Nice example, we would have said: 'If a truck appears after 9 p.m. and is within this 2-kilometer boundary of the boardwalk, we would have been able to have our software automatically flag this situation and notify police—via text, phone, or whatever,'” he adds. Giving the police that potentially actionable data opens up the possibility of stopping an attack—or at least lessening its impact.

The value of an IoT platform is the ability to convert meaningless data into meaningful and actionable information. “Ultimately, it is all about data correlation and data filtering that makes for the real value,” Chong explains. What we have been doing is to help to filter data and to manage and to correlate it in real time. That is what people need to make a decision and take action.”