Marinus Analytics Equips Law Enforcement to Fight Human Trafficking with Amazon Neptune

On the internet, human trafficking operates under a digital cover of darkness. The criminals involved have long had a technological advantage over law enforcement, using countless online outlets to exploit victims and reach across borders. The founders of Marinus Analytics saw that disparity and created Traffic Jam, a lifesaving tool that law enforcement can use to identify and take down even the most evasive criminal networks.

In 2019, Marinus Analytics saw success with its relational database infrastructure, identifying at least 3,800 trafficking victims—the majority in the United States—and detecting organized crime by surfacing relationships among data. However, link charting, or the process of visualizing relationships across entities, was manual and time consuming. To automate link charting and save hours of effort for law enforcement, the company needed a purpose-built database solution. To that end, Marinus Analytics sought out Amazon Web Services (AWS) to pair its relational database with a graph database, which can discover complex connections among data. Marinus Analytics chose Amazon Neptune, a purpose-built, high-performance graph database engine optimized for storing billions of relationships, to help it save law enforcement hours of manual effort in finding the biggest organized crime networks.

Cara Jones and Emily Kennedy founded Marinus Analytics in 2014 and created the company’s flagship application, Traffic Jam, to build on the work Kennedy started with her senior honors thesis at Carnegie Mellon University. Serving the United States, Canada, and the United Kingdom across all levels of law enforcement, Traffic Jam combs data from websites known to operate in human trafficking, relying on internal research and feedback from users globally to identify potential platforms and sources. “We created a repository of online classifieds resources where victims are often found,” says Jones, CEO of Marinus Analytics, “and that typically comes in the structure of a title and a body of text along with imagery, date, and location.” Those marketing snippets then become data points that users can tap into to find connections when queried and help investigate commercial sex activity.

Information discovery like this is critical because when police begin investigating a human trafficking case, they often know very little about the trafficking organization involved. The Traffic Jam app was built to put this interconnected information directly into the hands of law enforcement, allowing them to enter search terms—a name or an address, for example—and reveal any previously known connections.

Though Traffic Jam was able to capture approximately 40,000 points of data each day and help law enforcement achieve results, Marinus Analytics discerned that, due to jurisdictional boundaries, law enforcement needed still more investigatory help when it came to large multicity criminal organizations. Marinus Analytics wanted to query the database for more relationships within the vast field of data points to help law enforcement make connections within large criminal networks. But in a relational database, a user can query for only one degree of separation in data relationships at a time. To address this limitation, Marinus Analytics turned to Amazon Neptune for a more complex graph database solution.

Marinus Analytics already ran Traffic Jam on the AWS Cloud, benefiting from services like Amazon Rekognition, which uses artificial intelligence to apply facial recognition to photographs. When Marinus Analytics was searching for a graph database solution to complement its Amazon RDS relational database and optimize the Traffic Jam app, the company chose Amazon Neptune because the graph database can analyze data relationships, looking at nodes—or pieces of information—and the complex, web-like connections among them. Marinus Analytics engineering lead Ray Giorgi explains the connections enabled by Amazon Neptune: “We can start to bring in degrees of separation to show that ad A is within two degrees of separation of ad B because they share a phone number, or now ad A is within three degrees of separation of ad C because ads A and B share that phone number, and then ads B and C share an email address.”

AWS supported Marinus Analytics throughout the integration of Amazon Neptune in January 2020, and the company saw an immediate impact from the new graph database. Traffic Jam could now traverse relationships among data points and match those points across tens of thousands of ads with 2.5 degrees of separation, all while collecting approximately 350,000 data points every day. Amazon Neptune is also fully managed by AWS, which makes it easy for the 10-person team at Marinus Analytics to focus on Traffic Jam and its capabilities with minimal effort spent on managing its underlying Amazon Neptune cluster. Not to mention Amazon Neptune lends considerable speed; because of the graph database’s data access pattern, Amazon Neptune enables the application to query more efficiently than it can with a relational database. Now Traffic Jam can make those connections and return those queries in just a small fraction of the time it would take a relational database to perform the same query, which in most cases means a query takes milliseconds instead of minutes.

In the hands of end users, Traffic Jam is effective not only at giving law enforcement the contextual framework needed to approach multicity and multijurisdictional trafficking cases but also at querying the dataset “an order of magnitude faster with Amazon Neptune,” says Giorgi. Thanks in part to Traffic Jam’s new Amazon Neptune–powered graph database, law enforcement has discovered that some suspected trafficking groups have as many as 90 physical locations operating in the United States. Further, Amazon Neptune helps organize the massive amounts of data mined from websites, quickly connecting dots within these vast networks’ online activity and discovering actionable leads. This lets law enforcement spend more time doing their jobs and less time tied to a computer. “We want them to be able to go in and ask their questions, get the information that they need, and get back to combating criminal activity,” says Jones.

With the power of Amazon Neptune, Traffic Jam has been a huge game changer in empowering proactive police work, enabling law enforcement to investigate known or suspected human trafficking operations. “I don’t think that we would have been able to do what we’ve done without having Amazon Neptune’s tool-building environment,” says Jones. “It has been very empowering to take advantage of all the emerging capabilities for the social mission that we are very motivated to serve.” Marinus Analytics seeks to use its sophisticated architecture to go a step further and identify red flag indicators of trafficking: “We really want to keep pushing this proactive element and be able to lift out the largest connected subgraphs in our environment so we can reveal these organized crime networks,” says Jones.

Even though Marinus Analytics does not interact with victims, it serves those who do in a very real way. Thanks to the law enforcement and other agencies that use Traffic Jam, all this high-level data translates into lives rescued from the exploitation machine. Newly optimized, Traffic Jam is in use in the United States, Canada, and the United Kingdom across all levels of law enforcement, as well as by US Department of Justice task forces and nonprofit organizations such as the National Center for Missing and Exploited Children. These groups can now use Traffic Jam’s graph database to better drag human trafficking into the light, uncovering the enormous scale of the criminal organizations they are fighting.

To learn more about Amazon Neptune and other AWS purpose-built databases, visit aws.amazon.com/products/databases/.