With deaths from drug overdoses now the leading cause of death for US citizens under the age of 50, the US Department of Justice has turned to big data to help stem the flow of illegal drugs bought on the shadier back alleys of the internet.

This past August, Attorney General Jeff Sessions created a data science and analytics team called the Opioid Fraud and Abuse Detection Unit specifically to focus on sale of opioids online. Industry blog .pharmacy estimates that over $325 million worth of illegal opioids from China are shipped through post offices in the US. According to the CDC, overdoses from opioids killed about 66,000 people in the US in 2017. Some two million people in the US are addicted to opioids.

This January, Sessions doubled down as he ordered the Joint Criminal Opioid DarkNet enforcement (J-CODE) team to penetrate, infiltrate, and take down online opioid vendors. The FBI will detail dozens more special agents and intelligence analysts to assist the J-CODE team.

In addition to the effort to stop inflow of illegal drugs bought over the internet, the Opioid Abuse Unit will also use data science tools to sift through tens of millions of transactions reports from pharmacies to find out which doctors are writing too many opioid prescriptions in comparison to their peers and local pharmacies that are handing them out in greater numbers than their competition.

The Dark Net is made up of online sites and services that can be accessed only by special browsers that emphasize anonymity. Among the most popular of these browsers is Tor, which was formerly a fully-funded initiative of the US government. Tor was envisioned to provide security for vulnerable people whose national governments were intent on censoring the internet. Tor helped these people send emails and access information safely.

But the Dark Net also proved useful for criminal elements not only to conduct drug deals, but to deal guns, arrange contract killings, and distribute child pornography. When these deals were sealed with exchange of cryptocurrencies like Bitcoin, the participants seemed invulnerable to detection.

But trust data scientists to come up with solutions.

The Bitcoin Veil is Pierced

Last January, a team from Qatar University led by Husam Al Jawaheri found a way to connect social media posts to transactions on the bitcoin blockchain clearly enough to identify individuals. “The weakness that Jawaheri’s team identified is that bitcoin enthusiasts over the years have given their public keys out to lots of people and posted them in lots of different places. Many of them did so using internet accounts on which they also posted bits and pieces of identifying information about themselves, even if they wrote under a pseudonym—things like where they live, what they do for work, where they went to school or grew up, how many kids they have, whether they’re married…By scraping sites on the dark web for bitcoin addresses and then cross-referencing them with those same addresses on the blockchain, Jawaheri’s team was able to link 125 unique users to 20 Tor hidden services, “including sensitive ones, such as The Pirate Bay and Silk Road,” reports Mike Riggs on the reason.com website.

Laying Bait

Another data science-powered effort to penetrate drug deals on the Dark Web is led by information security firm Terbium Labs.

Alex Woodie from analytics firm Datanami writes that Terbium put juicy data bait out in the wild and waited to see who would show up. Terbium built “a substantial network of Web crawlers that are constantly searching the Dark Web. These crawlers cover an estimated 90 percent of the Dark Web domains accessible through the Tor network, including password protected websites, Terbium co-founder Michael Moore estimates. Because many of the criminals using Dark Web forums will share a sample of their stolen trove of data treasure to prove that it’s legitimate, Terbium is able to detect when a stolen credit number, for instance, is offered up for sale.”

Some observers say that despite recent triumphs like the takedown of drug marketplaces like AlphaBay and Silk Road, new marketplaces simply emerged to replace them. Some researchers reveal that there are about 32 more such markets operating in the Dark Net, just waiting for their opportunity to rise as the top markets are taken down.

Another aspect of the data science thrust in law enforcement is a renewed interest in reducing financial fraud. Unscrupulous futures traders were shaping the market through “spoof” trades. These spoof trades were huge orders for a futures contract that are placed but never executed. Such spoof trades are almost immediate cancelled but not before they roil the markets with misinformation. Such spoof trades were usually hard to identify amidst the daily trading volume but data science techniques have simplified the collection of evidence against erring traders.

The DOJ’s market data analysis has led to the filing of charges against eight people last January.