The Center for Data Innovation spoke with Dale Kunce, global lead for information and communications technology (ICT) and analytics at the American Red Cross, based in Washington, DC. Kunce discussed the value of open mapping data for disaster response and how the use of this data has reshaped humanitarian efforts around the world.
Joshua New: You help build Missing Maps, a project that mobilizes volunteers to map populations that are vulnerable to crisis or conflict. Why was there a need for Missing Maps? Why is this kind of granular, up-to-date geospatial data so important for humanitarian efforts?
Dale Kunce: Missing Maps was founded initially on a frustration. We had a couple large disasters where we didn’t have the data we needed to do our jobs as efficiently as we wanted.
There’s a small nongovernmental organization (NGO) called the Humanitarian OpenStreetMap Team that organizes digital volunteers immediately after disasters to do amazing work mapping disaster areas. Being the Red Cross, we thought we could do a better job if we prepared for what we think is going to happen so we can be ready for it. Missing Maps was founded because we thought we had the right skills to prepare for that next disaster—a typhoon in the Philippines, or an earthquake in Bangladesh—and, by tapping into our global network of volunteers, we knew we had the manpower to do it. So we started mapping areas we thought were highly disaster-prone or vulnerable. This could be natural disasters, or in the case of our partners like Médecins Sans Frontières (Doctors Without Borders), this could mean areas where malaria or yellow fever are endemic. Missing Maps is a partnership between all these different NGOs to help map the world. This is a very hard problem—just because you went to a place to map it doesn’t mean you’re finished, because you have to keep it up to date. I don’t think Missing Maps would have been possible without a collaboration between this consortium of NGOs.
New: Missing Maps relies on volunteers and crowdsourcing, which poses substantial challenges for ensuring data is accurate and complete. How do you account for this?
Kunce: The best crowdsourcing model out there is Wikipedia. 15 or 20 years ago, people laughed and said “there’s no way you can make an encyclopedia with crowdsourcing—that’s too much knowledge and you can’t trust regular people to do this job right.” I’ve been a geographic information systems (GIS) professional for around 15 years now, and I’ve learned that you can absolutely trust 1,000 people to make maps. We’re using OpenStreetMap, an openly licensed world map that’s been coined “the Wikipedia of maps.” Every time someone makes an edit someone else is following along and can verify it if it isn’t perfect. This means the maps are constantly improving, the same way that when major news breaks, it’s on Wikipedia immediately.
It’s really hard to do this an an entirely volunteer projects, but we’ve been successful at blending the knowledge and expertise of our own staff with that of the 19,000 volunteers that have contributed to Missing Maps just over the last couple years.
New: Your role used to be primarily GIS focused, but now you are focused on analytics more broadly. What other kind of data beyond GIS data are valuable for coordinating humanitarian response efforts?
Kunce: The first dataset we like to get is one that tells us where people are. With this, if there’s a hurricane for example, we could see how many people were in its wind path and plan accordingly. The challenge with this is that a lot of places don’t have an accurate census. In the United States we have a major one every 10 years, but some African countries haven’t had a census in 50 years. Other countries have censuses that are wildly inaccurate, and there are questions about whether or not you can trust the official government data.
We use a lot of the open data that we create, and then take in other datasets, such as data about a country’s growing season, which can tell us more about the damage a storm could cause. For example, with Hurricane Matthew earlier this year, Haiti wasn’t able to harvest crops before the storm hit, which contributed to food security issues. All of this data can tell us where people are and what we can expect the damage to like. In the first week after a disaster, we use this data to organize our plans. In the next three months, we apply this data to help address immediate needs and figure out how to best respond to long-term needs.
New: Can you explain the work the American Red Cross does to help develop open data standards in the humanitarian field? Why is this necessary?
Kunce: I’ve worked for many years in the private sector doing proprietary stuff, and I’ve worked for many years in the nonprofit sector doing open data stuff. There is simply no comparison to the value of having truly open data. It doesn’t matter what specific niche you’re working on, but as long as you’re publishing it in the open, anybody can help make value with it. By operating completely in the open, there’s no dark corner for us to shine a light on and find a magic piece of data—when all of it is open, we’re enabling use cases that we couldn’t possible imagine otherwise.
Just recently we were working on a project in west Africa and trying to map an area the size of Switzerland. We received a grant to map 7,000 villages in this area that had never been put on a map before and by going to each of these villages we built this really incredible dataset. The U.S. Centers for Disease Control and Prevention (CDC) heard what we were doing and now that our project is winding down, they are running with this data and will be doing all kinds of valuable analysis that my team doesn’t have the expertise or impetus to do. CDC is now generating value from our data that could eventually benefit us as well as the villages we helped to map. This same kind of thing could happen during disasters, where my team is focusing on mapping one half of a disaster area, while another team maps the other half. There’s no need to send them a file—we all put our data in the same spot as everyone else, on OpenStreetMap, and we both get the benefits immediately.
While operating off the same database like that is great, there are still times when we do need to transfer files. There have been a lot of great groups and tools developed in recent years that can help with this. One of these is called the Humanitarian Data Exchange, which we’ve been advising to help set humanitarian data standards and shape how we think about data to make everyone’s work a little bit easier.
By focusing on open data, we can make sure these maps belong to everyone. Our expertise goes into making the maps, but our goal isn’t to just go in and map a place. It’s to go and help people say what their place looks like. In Zimbabwe, we were working with the community to help them map what was important to them so they could use this data meaningfully. An older women suggested we map trash piles, which we initially thought was strange. It turns out the city wasn’t picking up trash as regularly as they should have, which led to huge piles of trash throughout the city that smelled bad and were making people sick. Right after, someone from the city government tried to convince us that we didn’t need to map the piles, but we went and mapped them anyway. Lo and behold, all that trash was quickly cleaned up and the garbage trucks now come more often.
New: You also previously led the global Red Cross network’s GIS analysis and coordination for the response to Typhoon Haiyan in the Philippines in 2013. How did this effort reshape humanitarian response?
Kunce: The precursor to Typhoon Haiyan was the Haitian earthquake in 2010, where 600 dedicated digital volunteers made something like 600,000 edits to OpenStreetMap. They built a map of Port-au-Prince that had never been seen before—if you look at before and after comparisons of maps, there used to be only one road, and then OpenStreetMap mapped a whole network of roads. A couple years go by and groups like the American Red Cross kept coming back to this example to build expertise around this kind of mapping.
With Typhoon Haiyan, the mapping community started mapping areas two days before the typhoon even hit in anticipation. When the storm did hit, within 48 hours this gave us a really good understanding of the road network and how the area was most likely to be impacted. In addition to that data, volunteers made something like 1.5 million additional edits in just the first month, which gave us a lot to work with. Primarily we used this to make paper road maps to distribute to the teams we were deploying so they would know how to get where they needed. We had these same maps on tablets that were being used to route workers while incorporating frequent updates. That allowed us to target our distribution for relief items so much better, and these maps even helped locals get around after damage from the storm changed things so much.
It also allowed us to see how groups of villages fit in with the road network. One of our relief team folks told me that the maps we were providing them cut a month off the time it would have normally taken to distribute their relief items. That’s a huge difference, and Typhoon Haiyan was the first to see this kind of huge effort built on OpenStreetMap. The U.S., Canadian, and Philippines governments were using it, we were using it, Doctors Without Borders was using it, and it became the definitive dataset.
You can continue the arc of this growth around open data. The Ebola outbreak happened around six months after Typhoon Haiyan, and 1,800 volunteers made around 4.5 million edits in just a few months to maps in west Africa that my team continues to work on years later. Then the Nepal earthquake in 2015 had 2,000 people make 6 million edits in the first 48 hours. We ended up having over 8,000 people contribute to open mapping this fall, and the amount of data available to humanitarians is growing on a logarithmic scale.
I heard a great story from our CEO Gail McGovern—she met with her counterpart in Nepal who told her that these maps and data were amazing, and that people were able to identify roads in Kathmandu that they didn’t even know existed. I think about the Red Cross as providing the help people need to overcome that next obstacle, and here we were having a huge impact on helping people help their neighbors. Sitting in Washington, DC, at a computer at midnight tracing buildings and roads, you don’t always feel that impact. But knowing that this data is being used so meaningfully is incredibly satisfying.