Using network visualization to deplatform an extremist group from social media
Tackle real-world challenges through advanced network visualization.
The Challenge
A multinational law enforcement agency (LEA) was tasked with disrupting the online recruitment efforts of an extremist group. Removing accounts one-by-one from their preferred social media platform was ineffective; the extremists quickly created new accounts and picked up where the deleted ones left off. To effectively disrupt this resilient operation, the agency needed to deplatform the entire network of extremist-affiliated accounts simultaneously. The key was to identify the extremist accounts so that a list for deletion could be provided to the social media platform.
However, doing so was like hunting for needles in a haystack. The extremists hid amongst millions of users on the social media platform. Unfortunately, the volume and complexity of data exceeded the capacity of the tools available to law enforcement personnel: the industry standard software simply crashed, while the high end platform ground to a halt. Dividing the data into smaller chunks was a non-starter, like trying to solve a 1,000 piece puzzle while working with only 10 pieces at a time.
Moreover, law enforcement officers with data scientist expertise and domain knowledge are in short supply, producing a major bottleneck in analysis capabilities.
Our Solution
Given these considerations, the organization chose Kineviz’ GraphXR. GraphXR’s backend graph architecture and in-memory capabilities meant that the agency could load all of the data, manipulate it, and visualize at the same time, and without the risk of fatal crashes. Its 3D visualizing capabilities allowed the analysts to examine all of their data at once, and dynamically, in order to see connections and relationships that would have been hidden by a 2D representation.
Its built-in data transforms and algorithms gave even novice analysts the ability to quickly make sense of their data and zero in on the indicators they were looking for. Finally, GraphXR offered an easy-to-learn environment that enabled the agency personnel to quickly get up to speed and to be productive.
These advantages made it possible for law enforcement to quickly review individual profiles, messages, and images in order to evaluate which were likely to be part of the network they were investigating.
The Results
GraphXR made it possible for law enforcement experts to leverage their knowledge of extremist networks, even without advanced data science backgrounds. This dramatically improved their efficiency.
The LEA successfully mapped the extremist network, a goal they’d pursued for years, in a period of months. The LEA delivered a list of accounts to the social media platform for simultaneous removal. As a result, in one day, the social media platform successfully deplatformed the entire extremist network. Subsequently, the LEA observed that the extremist group struggled to establish a foothold on other social media platforms, indicating that their propaganda and recruitment operation had suffered a severe disruption.