Big data companies are analyzing it and scientists are mining it. And now Sunandan Chakraborty is part of a team at New York University that’s using data to track the illegal wildlife trade online. He joins Hari Sreenivasan.
Using data to track the illegal wildlife trade
Big-data companies are analyzing it and scientists are mining it, and now Sunandan Chakraborty is part of a team at New York University that's using data to track illegal wildlife trade online.
He joins us to tell us more.
So, first let's get a basic definition of big data for people who are watching, especially in the category that you're working in.
So, as the name suggests, big data involves a lot of data, so now we are at the stage where we have access, thanks to the Web, huge amount of data in different form -- text, image, other-structure data, like weather data, transport data, et cetera.
So now we are at the place where we can use these different sources and different formats of data and can think of, can imagine doing things which was not possible before.
So, in terms of wildlife protection, what are data sets that you're able to access?
So, this project is mostly about detecting wildlife trade, and so we need to have data through which is about trade data.
So we are looking into places in the Web where things are bought and sold, like online marketplaces.
It can be online auction sites, or retailers having their own websites to sell stuff online.
So it's a collection of this type of site.
In addition, the modern trend is about using social media to trade.
So what's the E.G.I. Project, the Enforcement Gaps Interface?
What is that?
So it's a tool, a protected tool, whose job is to mine the Web as much as possible to collect all kinds of ads and postings from across the world in many different languages to identify all types of wildlife perhaps being sold online in different forms, and the main purpose of this tool is to find the subset of those which are not legal.
Okay, so give me an example of a product that you can find using this tool that you didn't know you were able to see before.
You find an ad, you try to understand what type of product it is, or whether it's ivory.
Sometimes sellers use code words.
So the job of the tool is to -- and the code word can be, say, ox bone or bovine bone.
So the job of the tool is to analyze the information element of the ad and match it with the larger set of data we already analyzed, or already have processed, and try to figure out, based on the price, based on the image, based on the text available, based on the item location, whether this particular ad or the product, which is being advertised, is ivory or not.
Or, similarly, if we can identify the species.
There are many species which you cannot trade internationally, that item cannot cross international border.
So we have a set of focus, a set of species, and we are trying to identify whether an ad, whether it's a taxidermy object or some item, which can be like boots, belts, coats, are coming from a species which you cannot trade legally.
And this would take a human millions of years to try to do what this algorithm can try to...
At this scale, absolutely.
Because there are, like, hundreds of sites, and for a human, most of the existing works have been done manually, and we are trying to make it better, make it like to work at large scale so that, ideally, we don't miss anything out there which is being sold.
What's the ripple effect?
What sort of an impact do you think that this can have, say, in the ivory trade?
So, first of all, let me start saying that this is only one piece of the puzzle.
We are only looking into what we call the Open Web, and then there is something called the Closed Web, which is like hidden behind password-protected sites or like WhatsApp accounts or e-mail accounts, which we don't have access to, and then there is the Dark Web.
So the thing is, what we have observed, this trade is so rampant in the Open Web, and so we are trying to make awareness, as well as detect these things in the Open Web.
So the one ripple effect which might happen is, people might -- Because it was so easy to do things in the Open Web, things might move a bit more into the Closed or the Dark Web.
Sunandan Chakraborty from NYU.
Thanks for joining us.
Thank you so much for having me here.