Mapping Catastrophe | Kent State University

Mapping Catastrophe

New GIS Techniques Generate Data in Disaster Zones

We recently sat down with Dr. Andrew Curtis, Professor of Geography and codirector of Kent State’s GIS Health and Hazards Lab. Curtis has spent the past several years developing new mapping technology that generates meaningful, realtime data in complicated environments. His research has given problem-solvers a new tool for helping communities overcome serious challenges at the local, regional, national and international levels, from Portage County to Cambodia.

HOW LONG HAVE YOU BEEN AT KENT STATE, AND WHAT HAVE YOU BEEN DOING?

About 5 ½ years. I came here from the University of Southern California.

When we first got to Kent, this room was just an open box. Part of my mission here has been to build this research lab and focus on different geospatial approaches, mostly to do with health but also to do with disaster as well.

And we have developed new techniques and new technologies. Even prior to coming here, we had the genesis of some of these, but it’s really over the last four or five years that we have evolved the methodology to the point that people in this country and abroad have found it useful to collect data that are not normally available.

WHAT IS THE SPECIFIC KIND OF DATA WE’RE TALKING ABOUT, AND HOW DO YOU GET TO IT?

Well, I work in what I call challenging environments, meaning it could be dangerous, or a post-disaster environment, or because there’s an ongoing epidemic. Alternatively, it could just be a data poor environment. So we’ve developed a couple of methods — spatial video and a spatial video geo-narrative — that allows us to create data where typically none exists. So a good example would be after a tornado. Say a community has been badly hit by this tornado, and if we want to monitor the damage, and then the recovery, these data layers don’t really exist. Now we collect data to create new maps based on how that community responds, how it rebounds after the disaster.

Alternatively, it could be the homeless, such as in the Skid Row area of Los Angeles. Thousands of people congregate in this one area, many living in shelters, or on the streets. The challenge here is we don’t have any data about them, which is problematic if a disease outbreak occurs within this cohort. So the spatial video approach can create maps where data is continually shifting in space and time —such as where the homeless camps are. By adding a recorder and then driving or walking with either an expert or somebody who lives in that environment and who can talk about what they see, or their experiences, now we have “context” to add to those maps.

Let me use an example closer to home. We want to support Portage County efforts to address the local opioid overdose situation. By interviewing the right individuals, whether they’re police officers, or health workers, we can learn about not only about where overdoses happen, but why and what is different over the last six months or a year. Now we can map that information in such a way that people who are not walking those streets, such as policy makers, can now understand the problem, and potentially develop intervention strategies which are based on insight.

It doesn’t matter whether we’re working in this country, doesn’t matter whether it’s mosquitos in a rural environment, overdoses in a small town, crime patterns or homelessness in a large city, this technique works in all these environments and across all different subject matter. Equally so, we can apply it overseas, where we have projects in Haiti, Ghana, Colombia, Cambodia, Rwanda, Guatemala, Zambia just to name a few -- all these different data-poor locations, which means we don’t have a lot of information on the ground. Now we can map out standing water where mosquitos might breed, trash concentrations that attract vermin, contamination points around public water access – all risk factors for disease, and then the narrative can help add further context.

One way to think about it is like this -- you want to walk in the shoes of with the most vulnerable. The spatial video geonarrative allows you to hear their voice and gain their perspectives on what is happening.

THINGS COME TO MIND LIKE A TENT CITY IN AKRON, HURRICANE RECOVERY IN PUERTO RICO, OR POLITICAL TURMOIL IN VENEZUELA, RWANDA, AND KENYA. CAN YOU ACTUALLY GET AN UNDERSTANDING BEYOND WHAT’S ON THE NEWS?

Exactly. It can be used by researchers, but also other organizations — whether they are service providers, health providers, non-profits, data is vital commodity for everyone to help understand the situation, and without it it’s hard to develop effective strategies. So, although we have not worked in Puerto Rico, we worked in post- Katrina New Orleans for a decade after the Hurricane. We were part of the response to the catastrophe, and then monitored the subsequent years of recovery.

What our post disaster work has shown us is that it’s a complex mess of emotions and experiences that shift back and forth from memories of the event to daily challenges faced in the recovery period. And these experiences can include perspectives on rebuilding, what the community would have liked to see happen, what did they think was missing from that process, all wrapped up with individual experiences and emotions that at any point could manifest into a health problem. So not only do we give a voice to those who are most impacted, but we can also read between their words and see how they are making observations. It’s not just a top-down imposition of “this is what’s best for you.”

The importance of hearing these different perspectives and voices extend to other settings. We worked in Skid Row for a couple of years where we were collecting homeless narratives. Initially this was part of a collaboration with the LA County Health Department, and especially TB control. The homeless suffer from a wide array of health problems —mental health, infectious diseases, chronic diseases, drug use and overdoses, while also experiencing almost daily violence— but we have very little data about this population. So, like I said before, if an outbreak occurs — a recent example being hepatitis, to develop an effective response we need to know how this world “works”. Then we can develop culturally appropriate, and therefore effective interventions. The narratives help give us this insight. For example, we can map out where overdoses are most likely to happen, and therefore where best to intervene. How to get Narcan, for example, into the right hands. And if we develop a health strategy, what is the most effective way to interact with this population so that they don’t see it as if it’s being imposed on them. So it’s basically a culturally-sensitive way to intervene, rather than just thinking “okay, there’s a problem, we have a solution, we’re going to impose it on you,” because that’s never really worked with a marginalized population.

TO WHAT DEGREE IS WHAT YOU DO DIFFERENT FROM GOOGLE STREETVIEW?

Google Streetview is a fantastic piece of technology, and there is now a lot of research that is using it as a data source. The limitations are, for example, when was that road driven? Did it cover all the areas you want? And sometimes there is a mix of different time frames, so if you’re not careful you could move between 2017 to 2014 to 2016 all within the same street segment. It’s a great source if you have no alternative data. However, if you are interested in how an environment has changed after an externality then it’s less useful. And, again, it doesn’t give you context. It gives you a visual, and the visual itself is certainly an improvement on traditional data. For example, if there are foreclosures, have these foreclosures turned into blighted properties? Do they turn into overgrown vegetation and graffiti? But with our approach we control when that visual is collected. And then with that narrative you can now find out what impact these blighted properties have on that community. How does blight affect the lives of the children? How does it affect access to the local park? And then, if there is an intervention you could again interview the individuals and say “have you noticed any change? Have things got better? Was removing this one house a real benefit to your community?” Because now what you’ve got is a template for elsewhere. You not only validate a strategy, you could say it works, and then translate it to other environments.

WHAT HAVE YOU SEEN BY WAY OF RESULTS? HAVE COMMUNITIES BEEN ABLE TO TAKE YOUR DATA, IMPLEMENT IT, AND FACILITATE SOME KIND OF REAL CHANGE? HAVE YOU BEEN ABLE TO DOCUMENT THAT CHANGE?

Yes. For example, St. Luke’s Foundation in Cleveland has had our camera kits for probably a year now, and they are using it to document various aspects of their community. And they send their data here, our students have helped with transcription and with mapping.

In another example, we have produced risk maps associated with standing water for public health use in Colombia and Nicaragua.

But maybe the clearest example came recently when we did a ride in one community with a police officer focused on overdoses. This officer commented on the amount of overdoses occurring in one particular restroom. Then the commissioner of public health, who was actually on that ride as well, said “well we need to start putting Narcan in that particular fast food place.” So that’s a pretty direct impact . Even if he had not been in the car, the software we have developed would still have let him see and hear those insights virtually. I just have to say that this last example is work I’m doing with Eric Jefferis in the College of Public Health. I’ve been lucky enough to work with him on several funded projects – it’s always vital to have a partner who is an expert in the substantive field. Luckily, we have an excellent talent pool of researchers and a culture of collaboration here at Kent.

WHAT’S THE KEY TO MAKING THIS RESEARCH AND THESE METHODOLOGIES SUCCESSFUL?

For our work to be successful, it has to be used by everybody. So part of what we have developed here are new software so that the data we collect can be used by a community group, a police officer, by a public health worker, and it can be used in Cambodia and Haiti. We have developed ways that the data can be transferred easily as well, and we’ve even developed a simple mapping alternative to GIS based on data that the user is seeing in the videos or hearing in the narratives. It’s vital to take this method outside of our research world, outside of academia, and give it to our collaborating partners, many of whom do not have the spatial technology or research method background that we have. We need to be able to translate the method, the data, the findings, in such a way that individuals who don’t have an academic background can still use it. We don’t want to just tell people what we find, we want to give it to them so that they can also use it and turn it into other knowledge.

HOW HAS YOUR WORK PROGRESSED FROM YOU THE RESEARCHER IDENTIFYING AREAS THAT YOU WANTED TO STUDY TO PEOPLE SAYING “WE’D LIKE YOU TO COME TAKE A LOOK AT OUR CITY?” HAVE YOU GOTTEN TO THAT POINT YET?

Absolutely. When we first started, it was in response to Hurricane Katrina. From there we took it into other disasters because we had an interest, we wanted to study how these environments were different to Katrina. From there, we started to develop research avenues into the health realm. At the same time, we started to get people who were aware of what we were working on, coming to us and saying, “can we use this in Skid Row, or can we use this in Haiti? Now we’re at a level, because of the software that my doctoral student Jay (Jayakrishnan Ajayakunar) has developed — which really has opened other possibilities, through this fantastic software that he’s coded — now we have people who come to us (like Portage County Public Health, or Akron Children’s Hospital, or Case Western Reserve University, or the University of California at San Francisco, or Stanford University) and say “let’s have a look at this situation” and we follow their lead. So most of the work we are doing now, I would probably say, has not come from me. It’s actually been our lab and our students responding to some outside request or idea about a collaboration potential. One of my doctoral students (Sandra Bempah) is working in Ghana this Christmas with collaborators there. Another post-doc (Amy Krystosik) who got her PhD from the College of Public Health here at Kent State continues to work with us in Colombia and will soon take these technologies to Kenya. In both these situations it was actually our students who drove the idea and collaboration. We are lucky to have such good graduate students either in Geography, or from the College of Public Health. These guys love this type of work as it puts them in the field – for example my masters student Rob Squires went with me to Haiti this January. But even in the classes I teach, because I expect a final project using these sort of technologies, the students have always responded because they love the idea of real world research. Students from these classes have gone on to become my graduate students, or just use these approaches in their own research or real-world employment when it comes along.

But returning to the idea about others seeing the potential; here’s a more local example. I’ve given a lot of talks across our region and these are often followed by people who contact me and say “that’s interesting; is there any chance we could do this?” After one such recent talk, I had a doctor from Akron Children’s Hospital — who are great partners we’ve worked with a lot over the years -- come to the lab and say “we’d really like to use this technique to start to look at community health, and start to see, where we’re involved, what is actually going on.” It’s always cool when that happens.

HOW COULD THE METHODOLOGY BE APPLIED FOR OTHER ACADEMIC OR EVEN REAL-WORLD DISCIPLINES? SAY HISTORY OR ANTHROPOLOGY OR JOURNALISM?

To give you a good example of more historical work we’ve been doing recently with Professor Tyner, over this last year we went to Cambodia, where we interviewed individuals who were part of the evacuation of Phnom Penh during the Khmer Rouge rule. Now although previous research has considered the evacuation, we actually followed the evacuation paths, the routes, with some of these individuals. And along that pathway there was insight, in terms of not only how the environment has changed, but also what was it like to evacuate along a particular road. What was it like to be at this location, on this bridge? And that’s a good example of how places, or spaces trigger memories. If I was sitting across the table and you were telling me about this, it would be a more generic conversation; there would be key moments on that path that you would talk about because they’re in your memory, but when you see a building, when you see a river, when you see a marsh, when you see a particular bridge, then your memory kicks in in a whole other way.

The virtual drive through is something we’d like to explore in the future, and which is possible because of the expertise we have here at Kent in computer science, especially Professor Zhao. So instead of taking somebody out into the field — that’s always going to be the best, because it’s not just the sights, but the sounds and the smells that stimulate — but for somebody who lives in Cleveland now, who was also part of the evacuation of Phnom Penh, it would be great to be able to put them in a lab like this and show them the different route-ways out, so that they can see it and talk about it, even though they’re never going to go back to Cambodia. This will be an improvement on a typical interview in a sterile environment, because we are at least giving visual cues. And in those visual cues, because the video is tied to the map, we can have a more accurate recording and mapping of where they’re talking about.

POSTED: Wednesday, April 11, 2018 - 3:10pm
UPDATED: Friday, May 11, 2018 - 10:46am