From unmasking Banksy to catching terrorists: the power of geographic profiling

Using math  — specifically geographic profiling — researchers Dr. Steven Le Comber and Dr. Kim Rossmo were able to narrow down the possible identities of the elusive graffiti artist Banksy. 

Dr. Rossmo speaks with Brent about the power of these geographic profiling methods and their possible use for tracking down terrorists and fighting mosquito-borne viruses like Zika and malaria.  

Their study, published in the Journal of Spatial Science, explains how they used the geotagged locations of Banksy's artwork to pinpoint the possible movements and likely home base of the anonymous street artist.

Banksy has gained a great deal of notoriety for his artwork, despite remaining unknown to the public.  

This interview has been edited for clarity and length.

Brent Bambury: Kim, what have you uncovered that suggests that Banksy may well be Robin Gunningham?

Kim Rossmo: Well, when we decided to do the analysis of the Banksy art sites, we originally were going to assess all the credible Banksy suspects. It turns out there's only one, which is Robin Gunningham. When we looked at the locations where the Banksy graffiti art is located, we found a very close match in terms of an analysis of those locations, and those places like home and school where Robin Gunningham has connections.

BB: So what were you analyzing when you were looking at the locations of the art, and the places where Robin Gunningham went to school or played soccer? Can you break down how those things related to each other in a way that gave you information?

KR: Well, when people choose to do something, whether it's a criminal committing a crime or a graffiti artist using a location to stencil something on the wall, they want to operate in areas that are not too far from their home; within their comfort zone. But they also don't want to do it right outside their door. So there's a relationship, and some balance, between being close but not too close. And while you can't say much from a given location of just one incident, when you're dealing with dozens — or in this case, 140 — the rules of probability are such that you can say something quite strong.

BB: Imagine that you're explaining it to a classroom full of fourth graders. How would you explain the way that your analysis works?

KR: We use a software system that takes into account the actual pattern of locations. So think of these as points on a map. That pattern is, in effect, a clue. And we can mathematically analyze that to try to determine the origin for those points. So whether it's a criminal leaving their home and going out to commit a crime, or someone leaving their home to go out and do some graffiti, we can try to determine the most likely origin for that particular pattern. And in this case, that would be where Banksy lived, worked and played.

BB: It sounds like there is some reasonable precision to the tools that you're using. What other mysteries could geographic profiling help solve?

KR: It was originally developed, when I was a doctoral student at Simon Fraser University as a policing tool — something that investigators would use if they were doing an investigation of, say, a rapist, or a serial murderer or arsonist. But since that time, we've found other applications in biology, zoology, epidemiology and counter-terrorism. One of the more exciting was applying it to epidemiology. We looked at cases of malaria in Cairo, Egypt. And then we looked at where mosquitoes that carried the malaria virus were breeding. And we were able to infer the location of the mosquito breeding pools from where the malaria cases had occurred. The best way to think about this is an information management tool, or a suspect prioritization tool. And the smaller the focus, the better the job we've done, because we've said 'Hey, rather than look at a thousand suspects, you can find who you're looking for in the top ten.'

BB: Serial killers generally operate locally. You tracked the Banksy data in a couple of cities. But terrorists operate globally. How do you zero in on a terrorist cell that might have components all over the world?

KR: Let me give you an example of an actual case we studied, which involved the assassination of a former minister of justice in Ankara, Turkey. This is a fairly complicated process; it requires a lot of setup. So the strings of the operation — the brains behind it — actually were some people in the Netherlands. They directed what they call a 'revolutionary propaganda team,' an assassination team from Istanbul. But they just don't show up one day and shoot the Minister of Justice. They have to set up a whole terrorist cell, which requires support. So they found places to live. They had to have storage places for weapons, which had to be delivered to them in the first instance. They had communication points. One of the people had to have a job. Because this took several months. And they needed places to flee to afterwards.

So when you take a look at the cell sites, the safe houses, and all these support places, you find a pattern very similar to what we see with with a crime. Now, the idea we got from that is, when you go to certain European and Asian cities there's often a lot of low-level terrorism activity by certain groups. For example, anti-government or seditious graffiti. Banners. Leaflets. Even low level crimes like shoplifting and credit card fraud, which are used to support the daily operations of terrorists. So the idea was — rather than wait for a big explosion from a bomb, and then respond — try to use the special information from all this minor stuff to then help direct where the focus should be.

BB: Has that worked for you, Kim? Have you been able to say that this has paid off, that you've actually prevented some attacks?

KR: Well, I can't tell you what the intelligence communities or law enforcement are doing. But I can tell you, in terms of our research, that we found that this technique works. We're less interested in identifying who Banksy might be than we are in showing how analysis of graffiti can focus on where the origin for that graffiti came from.

BB: In the case of Banksy, the graffiti was the what led you to the information; in the case of potential terrorists, it could be leaflets or other kinds of seditious graffiti. You also have a historical example involving Nazi propaganda postcards. Can you tell me how that worked?

KR:  Back in the early 1940s, a couple by the name of Hampel turned against the Nazi government. They began to leave postcards all over Berlin saying 'Don't trust Hitler, don't trust Goebbels.' And they would leave these postcards on the street, or in the lobbies of apartment buildings. And the Gestapo, with their usual efficiency, collected all these cards and analyzed them. They had a list of the addresses and the dates. So we were able to get that information and we were able to translate it. We went to Berlin and geo-coded all the sites, and then — because we knew where the Hampels lived, where their apartment was — we were able to see how accurately we could determine their home, based on the locations of the postcards. And it was extremely accurate.

BB: I want to come back to Banksy for the end of the interview. Banksy is known for satirizing the war on terror and surveillance. Do you think that there's an irony here — given that he's obsessed with intrusion and he's potentially been unmasked based on the pubs he visited and where he played football?

KR: There definitely is some irony. But I'm not sure how much privacy somebody really wants, when they're doing graffiti on public buildings.

BB: Well, but he's been obsessed with being anonymous.

KR: But there are lots of ways to be anonymous without attracting so much attention to yourself. I'd like to say, I like Banksy's art and I respect his social commentary. And we do need to have these dialogues, and we need to strike the proper balance between privacy and the need to protect ourselves.

BB: But where does your technology fit in with this? Do you worry that the technology you've developed could be used for nefarious purposes?

KR: Well, I guess anything is possible. But we've been doing this since the early 1990s, so I'm not too worried about it. Because all we really do is prioritize suspects and I would rather see an evidence-based, scientific method of prioritizing suspects than something based on a bias or a prejudice.