As the humanitarian response to last Saturday’s deadly earthquake in Haiti continues, responders face a seminal question: Where is aid most needed? Most often, the answer depends on where the most affected people are located. But in the aftermath of a disaster, especially as digital and transportation networks are reduced, collecting actionable data hasn’t always been a given.
But that has changed, according to Andrew Schroeder, vice president of research and analysis at Direct Relief and co-founder of CrisisReady, a collaboration between Harvard University and Direct Relief.
“Big data analysis has gotten much faster,” Schroeder said, noting that, within days of the earthquake, he has been able to get high-resolution satellite imagery, social media-based data showing population movement, damage assessments from the EU, geocoded news updates, and a landslide risk analysis.
With these resources, Schroeder and his team have created maps with a plethora of data, enabling response organizations and coordinating groups to deploy their resources to areas where they’ll have the most significant impact.
In Haiti currently, Schroeder has found that most people who left their homes in southwest Haiti, the earthquake’s epicenter, have stayed close to home and not traveled to Port-au-Prince or other big cities.
“What you see is that it tracks very localized movement. So far in Les Cayes, we can detect that there has been some level of a shift from downtown on the waterfront back into areas that ring the city and then into Camp-Perrin. But no farther, not to Port-au-Prince,” he said. “To the degree people are displaced, they’re displaced close to where they live.”
However, Schroeder cautioned that the sample size in southwestern Haiti is smaller than is typical for the Facebook-derived data upon which he based his analysis. Direct Relief has access to data through the Facebook Data for Good program, which shows anonymized Facebook app user movement after disasters.
This pattern stands in contrast to some natural disasters, such as Hurricane Dorian in the Bahamas, where residents of the Abaco Islands and other mostly destroyed places left for unaffected islands and major population centers.
Yet, according to Schroeder, a pattern similar to what is happening in Haiti occurred after Northern California’s Camp Fire, when many evacuees stayed near the destroyed town of Paradise.
Since so many social, logistic, and even psychological factors determine how populations move after a disaster, having access to real-time – or close to real-time – data is critically important to response efforts. Still, knowing where to go is only part of the mission.
“The trick is to get assistance to them. On [UN-coordinated] calls now, we hear that roads are unsafe, there was a temporary obstruction of roads between Les Cayes and Jérémie, there’s all kind of issues on the transportation network. This confirms what we know anecdotally,” he said.
Disaster Relief 2.0
Direct Relief’s data on Haiti is sourced from several partners. In addition to Facebook, Maxar Technology provided high-resolution satellite imagery, the European Union provided damage assessments via their Copernicus Earth Observation program, Factal provided geocoded news alerts, and Esri provided topographic information and baseline population figures.
Schroeder and his team have overlayed these data on a GIS map, thereby giving governments, international agencies and local responders improved situational awareness on how to best help people in the region.
An early example of this has been Facebook data on network connectivity, which was analyzed by the CrisisReady team and shared with NetHope, an NGO focused on restoring – and expanding – access to communications networks.
The movement towards collecting and sharing this data was catalyzed after the 2010 Haiti earthquake in what a 2011 report from the UN Office for the Coordination of Humanitarian Affairs called “Disaster Relief 2.0.”
“Crowdsourcing, drones, humanitarian open street maps – all the various things that came together in Haiti in 2010,” Schroeder said, referring to some of the specific technologies that comprise the increased data sources for disaster relief.
And while data collection and sharing have improved over the past decade, Schroeder said that some of the old problems still face the disaster response community, such as overlapping missions and a lack of efficient collective action.
As the response to Haiti continues, Schroeder hopes that technology will continue to improve efficiency and coordination, both for future disasters and in everyday humanitarian aid. He mentioned that Balcony, an app that connects people in an organization, has delivered promising results.
“With the app, location services are turned on so they can see where their people are, communicate, collect data, incorporate data into what they are doing, site assessment, data assessment, material tracking, that kind of digital coordination structure, there’s a plan to set that up,” he said.
In Mexico, the government has used this kind of tech to help coordinate their response to homelessness, with thousands of caseworkers logging each interaction, thereby enabling a map showing where those most in need of services are located in cities.
“There’s not a functional difference between that and the emergency response. Emergency response is just more and faster, and things are more broken,” he said.
Regarding implanting this kind of technology in the coming days and weeks in Haiti and during the next disaster, Schroeder is hopeful.
“We have the ability to do this,” he said.
For the latest on Direct Relief’s response to the earthquake in Haiti and its support for St. Luke’s Hospital and other Haiti-based organizations, visit directrelief.org/emergency/haiti-earthquake-2021/