The Center for Data Innovation spoke with Noel Dickover, technical director of global network strategies at PeaceTech Lab, a nongovernmental organization (NGO) based in Washington, D.C. Dickover discussed the value of getting data into the hands of humanitarian workers on the ground, rather than just the leadership, as well as the importance of making this data usable in combating violent extremism.
Joshua New: PeaceTech Lab was created by the U.S. Institute of Peace (USIP), an independent federal institution, to use data and technology to help prevent and mitigate violent conflict. Why was there a need to improve how data is used in this space?
Noel Dickover: USIP created PeaceTech Lab to apply media, technology, and data in preventing violent conflict and promoting secure and peaceful societies around the world. Starting in 2008, the Lab’s work began as a Center of Innovation within USIP. Now an independent non-profit organization, the Lab shares USIP’s physical space and benefits from the Institute’s 30-years of expertise in peacebuilding, while pioneering new ways to achieve the Lab’s own mission. The PeaceTech Lab provides tech, media, and data tools and training to peacebuilders and their allies to enable them to be more effective. We work in conflict zones to build and sustain networks of local technologists and peacebuilders interested in addressing violent conflict. The data focus allows us to engage in two-way information sharing relationships with these folks on the ground. From a situational awareness standpoint, the hope is to combine a big data view with a local network and local data to get far more granular in our understanding of the drivers of conflict, as well as the mitigation strategies that would be most effective in addressing the conflict.
Just as an example, the Lab is experimenting with ways to use big data to revolutionize how the needs assessment process is conducted for peacebuilding interventions. Currently, peacebuilding programs incorporate conflict analysis and needs assessment to determine how to prioritize resources and design the intervention. This involves a combination of desk research, outreach to partners, and on-the-ground data collection. While there are a bevy of new options to enable more efficient means of local data collection including in challenging environments, very few are exploring the possibilities of applying tools like social media listening and news analytics to build a comprehensive picture of the drivers of conflict before ever setting foot on the ground. These tools have the potential to support more targeted local data collection processes and can serve as valuable proxy indicators for ongoing monitoring and evaluation by measuring events and perceptions.
New: In March 2016, PeaceTech Lab launched the Open Situation Room Exchange (OSRx), which you head, to inform on-the-ground efforts in conflict zones to combat violent extremism. How does OSRx work?
Dickover: The “who is it for” question is the critical one. If you look at the investments in data systems present in this space, most are what I affectionately refer to as “data for superman” systems. The idea behind them is you have this really important person—it could be Obama, Secretary Kerry, Ban Ki Moon (frankly the list of important people go a mile long)—if we can give these important people this amazing situation room, which is almost always protected behind a firewall (a closed situation room if you will), in which we can provide everything they need, in the way they need, at the time they need it, the thought is “superman” will have “magical levers” at their disposal that they can use to, say, get rid of the problems affecting northern Nigeria.
But what if we don’t really believe superman has those levers? What if instead we think that the way northern Nigeria changes will be when the local advocates for peace can convince the population in the communities surrounding them to think about using nonviolent means for resolving their differences? If that’s how we think change happens in conflict affected areas, it’s really those local peacebuilders who need this amazing situation room to help them think about information in the world around them, and how it might be used for strategic planning or tactical awareness purposes. Bottom line, we can provide superman with early warning systems, but if we care about responding, we’re almost always talking to local advocates for peace.
We’re generally not building for people in these countries looking at 3D visualizations on 24-inch monitors. Far more often they have feature phones working in partially connected environments, where trust and information sharing is not freely given. This is why to the extent possible, data access on the OSRx will always be anonymously accessible.
We released the first version of the OSRx a few months back. We are about one year into what we see as at least a five year process of building out capability. Right now the OSRx has a global outlook page to compare countries and see incidents related to conflict around the globe, and country pages to dive into the details on individual countries. Data comes from event reports of protests and violent conflicts, underlying structural data, like the World Bank Corruption Index, for instance, and near real-time data from social media and news aggregation. Right now, we’ve parsed out global datasets to each of the country pages. Over time, as we connect local networks of peacebuilders to the OSRx, most of the data on the country pages will come from country-specific datasets, and the content on the pages will come from networks of peacebuilders within the country.
New: A big focus of OSRx is making data usable, rather than just dumping information. What does this entail exactly, and what are the challenges in making this data usable?
Dickover: The real challenge in the peacebuilding field is that because the drivers of conflict are so deep and intertwined—things like fragile governance, inter-ethnic and religious tensions aggravated by hate speech, gender-based violence, violent extremist activity, endemic corruption, and so on—there is a tendency to use “thick description” to describe the complexities involved. There is simply not a history of using data or visualizations as key to strategic planning and course correction on movements or interventions. There are very few data scientists in the space and most don’t have the time or knowledge to think through how this information might be applied.
It turns out there are a number of amazing open datasets online that, with some pre-processing, might create a new capability or knowledge resource that suddenly becomes meaningful. Our focus is to do the pre-processing to make this information more relevant to the problems and challenges peacebuilders face. Unlike many fields, few in the peacebuilding field are clamouring for data as a silver bullet—the challenge is quite the opposite—we need to really work with them and think through ways to make this information relevant.
Different user groups will have different access and knowledge needs. Right now, the current build of the OSRx has utility for researchers and policymakers. We still have a lot of work to do to make it usable for peacebuilders in partially connected environments, where risks from authoritarian regimes and nonstate actors really impact the options for information sharing. Near term upgrades include making it multi-language and optimizing for mobile use.
For instance, the Armed Conflict Location and Events Database, affectionately known as ACLED, provides a weekly updated Excel file of a protests and violent events across Africa and Asia. We take this information and place it on a map, and serve it up on all the country pages in Africa. Right below that, we have visualizations created from a a social media listening platform that allows the user to access the visualizations to explore the online conversation and media related to those violent incidents and protests.
So for instance, on March 7 in Tunisia, in the city of Ben Guerdane, there was a horrific attack that left over 50 people dead. If you looked on the Tunisia page of the OSRx in March, you would see in the social media visualization that the hashtag, “#BenGuerdane” was where the conversation in social media had coalesced. We have a Twitter search right below the visualization that allows users, even if they aren’t familiar with Twitter, to now access the full conversation around these events.
But really, the plan over time is to use formative usability studies and iterative design to find ways to connect the data to the activities involved in peacebuilding. As a minimum, it’s already clear to us that providing “dashboards of data’ is not going to get it done. Because people in this space really work with “thick description,” our current thinking is link the data visualizations to underlying narratives that describe the drivers of conflict. So if I’m addressing a “culture of corruption,” for instance, we are looking at blended automated and manual methods to tie a thick description narrative to the data. This is a key area of experimentation for us.
We also have information that can help put this violent event in context with the rest of events in the country. We’ve worked with Kalev Leetaru, who runs the Global Database of Events, Language and Tone, one of the largest open datasets in the world focused on worldwide news aggregation, to create an instability timeline to assign sentiment to the world’s news as it describes Tunisia over the past 180 days. You can see the impact of the Ben Guerdane attack on the country in comparison to other unstable events.
New: Before your role at PeaceTech Lab, you co-founded CrisisCommons, a network of tech-minded volunteers that leveraged technology and open data to aid disaster response and humanitarian relief efforts. How does open data play a role in these efforts?
Dickover: In the development space, it’s fair to say that disaster response and humanitarian relief has been the trailblazer on using open data in innovative ways. Haiti was a sea change in that sense, as most of the traditional responders were in the large buildings that collapsed. A myriad of nontraditional actors entered the humanitarian space, including volunteer technologists in loosely knit organizations like CrisisCommons. They applied a variety of data sources, both crowdsourced and existing datasets to help first responders in their relief efforts. Patrick Meier’s book Digital Humanitarians is probably the best source for understanding this evolution, which now includes remote sensing approaches linked with crowdsourced and machine learning methods for providing awareness data. Really interesting approaches like the Humanitarian Data Exchange platform and the Digital Humanitarians Network have transformed the volunteer technology community to a real force in the humanitarian space. The peacebuilding field is far from this in terms of maturity, but many of the same innovations apply. My favorite example of this is a project in Myanmar where we connected a wonderful machine learning company called Bindez with civil society organization, the Yangon School of Political Science, to come up with an approach to find instances of hate speech online to serve up to organizations looking to address the problem.
New: You also worked for several years in the U.S. State Department’s Office of eDiplomacy, and have a long private sector career. How has the public sector and NGO approach to humanitarian response changed over the years in terms of how they leverage data and new technologies? Are there any major lessons they can learn from the private sector?
Dickover: The key lesson from the private sector that applies is the importance of user experience and incorporating feedback from users in the design process. Cloud-based processing tied to mobile devices, even with just text messaging, is transformative for information collection and situational awareness. The context is really different though. Silicon Valley assumes ubiquitous connectivity and power in designing solutions. These are not givens in our world. Designing for partially connected environments is significantly different. This doesn’t mean we cannot use cool things like the Internet of Things for data collection, just that we have to really think through the details. William Gibson’s quote applies in the peacebuilding space: “The future is already here—it’s just not very evenly distributed.”