Online Child Sex Trafficking as a Data Mining Problem

Making Sense of Big Data in the Real Business World

One of the most frustrating Internet crimes is child sex trafficking, and in many ways, it’s a data problem.

I remember reading an in-depth report on this issue years ago. Police had all these photos of children being abused — in fact, one little girl appeared so often, police were literally watching her grow up.  They’d given her a nickname and despite their best efforts at visual analysis, they’d never been able to find anything to identify either the girl or the photo locations. It’s still one of the most heart-breaking stories I’ve ever heard.

The problem was they only had photos of the children. They wouldn’t risk publishing the photos of these exploited minors both for legal reasons and because they feared the children might be killed by their captors.

When we talk about Big Data, we tend to focus on volume, because that’s the most common use case, followed by velocity. But the final “V” of the trio is “variety,” which includes non-structured types of data.

Frankly, even though we don’t talk about “variety” a lot, some of the most exciting, cutting-edge Big Data use cases focus on photos and other image data. We talk a lot about health care and utilities, but Big Data technologies could literally revolutionize the fight to stop Internet child pornography and sex trafficking.

A recent Nextgov article highlighted three new technology tools developed to help victims of trafficking.

The article doesn’t specify that these are Big Data mining tools, but I’m willing to bet $1 that they use Big Data technology, since two of the tools focus on photo analysis to help identify the children being exploited in online ads.

One tool, developed by the nonprofit Thorn, relies on automated analysis of online ads. The other is a photo manipulation tool being developed by IBM, MIT and the Royal College of Surgeons in Ireland.

The third allows trafficking victims to send a text message to the National Human Trafficking Resource Center hotline.

It turns out that governments aren’t the only organizations interested in using Big Data tools for crime fighting. Security is one of the top five enterprise use cases for Big Data technologies, according to IBM’s Product Marketing for Information Management Vice President Eric Sall.

You may have known companies are using Big Data analytics in IT security and fraud prevention, but it’s also being deployed for improving perimeter security. Companies use Big Data to analyze video footage and identify suspicious activity. It’s also being used to analyze machine logs to detect analogies that may indicate and intruder, Sall told Silicon Angle.

IBM says the other major Big Data use cases for enterprises are:

• Big Data exploration
• Achieving a 360-degree view of customers
• Operations analytics
• Data warehouse augmentation