Combating Human Trafficking Using Data Science

There are more than 4.5 million victims of sex trafficking worldwide, according to International Labour Organization estimates. In the United States, some trafficking victims are often hidden in plain sight, forced into sexual exploitation and held captive in establishments disguised as legitimate massage businesses. There are around 9,000 illicit massage businesses (IMB) operating across the country, complete with websites, customer reviews, licenses and tax records. Despite efforts to shut down individual IMBs, they often reopen under different names, suggesting a deeper connection linking establishments, which was not always evident from the data.

Polaris aims to dismantle these trafficking networks. As a nonprofit, non-governmental organization that works to combat and prevent modern-day slavery and human trafficking, Polaris works directly with victims, engages systems to disrupt trafficking, hosts tip and crisis hotlines and offers support.

Polaris hypothesized that the resilience of the illegal massage businesses—the way they are able to open up in a new location after closing in another—indicates they are part of well-organized criminal networks; they were not simply “mom and pop” shops.

But without proof from the existing data—or strategies to visualize complex networks of IMBs—law enforcement agencies would continue to lose the fight against human trafficking.

To test the theory, Polaris tackled the human trafficking problem in a small city, manually identifying 35 illicit massage businesses and revealing how they were all connected. It was a difficult and labor-intensive process that relied heavily on cost-prohibitive tools. The answer and approach was right at their fingertips, but it wasn’t scalable—law enforcement, government, and prosecutors could not accommodate the manual labor required to handle and understand the data and connections between the information that could help shut down the illicit activities permanently.