In August 2012, Manila was flooded by an unnamed storm. Concerned citizens set up a Google spreadsheet, along with the hashtag #RescuePH, to crowdsource rescue needs across the city. One plea read, “Please help urgently. Mother named Cindy with kids and baby. Already on roof begging for help.” The 2012 spreadsheet was soon filled with hundreds of such requests, inputted directly into the spreadsheet or culled from Twitter and Facebook.
It was an early sign that, in the age of social media, citizens naturally gravitate toward crowdsourcing disaster relief, and institutions from city governments to Google are taking notice. How well such techniques work, and how they should be used, however, remains an open question. According to Anthony Abustan, a researcher at the Disaster Risk Reduction office in Marikina, a district along one of Manila’s most inundated rivers, tools like Twitter and Facebook may help assuage victims’ helplessness, but as of this moment, it’s “very difficult to verify whether the report is accurate or just a hoax. There’s just no way for us.”
Currently, Marikina’s disaster relief relies mainly on 100 CCTV cameras monitoring its area of responsibility, patrollers roaming the city on boats and ambulances, a rescue hotline and switchboard, and telephone reports from trained local leaders. In the last couple of years, disaster command centers have proliferated in Manila’s districts as well. Now the question is, how can systems like these be bolstered by a crowdsourced response?
While a simple spreadsheet might not be much use, a more comprehensive approach holds promise. Since the 2010 earthquake in Haiti, Google has been using its reach and scale to aid in crisis response. One of Google’s flagship crisis products is the Google Person Finder, a simple-to-use tool that allows anyone to input and search for the names of survivors. After Typhoon Haiyan, the number of records reached 60,000. Person Finder is widely accessible, and it can be embedded into any website, including the Philippine Department of Social Welfare and Development, which coordinated much of the Typhoon Haiyan’s rescue efforts. Names can be inputted or searched by SMS, which is crucial when internet infrastructure is down in a disaster zone.
In each of the major disasters since the advent of the social media age — the 2010 earthquake in Haiti, the Japanese tsunami in 2011, Hurricane Sandy in 2012, and 2013’s Super Typhoon Haiyan in the Philippines — survivors took to social media in droves.
“Welcome,” says Patrick Meier, director of social innovation at the Qatar Computer Research Institute, “to the world of big data, in which disaster-affected locations are increasingly becoming digital communities.”
According Meier, who specializes in crisis mapping, the Japanese tsunami generated at its peak more than 300,000 tweets per minute. Hurricane Sandy elicited more than 20 million tweets. In Haiti, where survivors used text messages to report immediate rescue needs, like people buried in rubble, Meier, working with an organization called Ushahidi, quickly came up with a system to manage the avalanche of information. Tragically, Meier said, “the backlog of unprocessed text messages grew larger with every passing day.”
Despite false starts and unresolved challenges, there’s no doubt that there’s opportunity in crowdsourcing information after a disaster. For rescue organizations, gathering information quickly is a matter of life and death in the first hours and days.
Meier is working on creating algorithms to help sort through the noise. His team researched the operational value of Twitter after a disaster, and after conversations with the United Nations Office for the Coordination of Humanitarian Affairs, they created a “Twitter Dashboard for Disaster Response,” which uses algorithms to identify relevant tweets during crises, teasing out infrastructure-damage assessments, casualties, humanitarian needs and offers of help, depending on the nature of the disaster.
Crowdsourced mapping seems especially promising. In the altered landscape of the Philippines after Typhoon Haiyan, the American Red Cross and other aid agencies used OpenStreetMap to create an open-source map within three days of the disaster using at least 400 volunteers the world over. The map helped humanitarian organizations navigate the buried roads and debris-strewn villages in order to deliver aid.
Haiyan’s OpenStreetMap generated an estimated four million edits from volunteers, an enthusiasm to help that’s been repeated over the past several weeks during the Ebola outbreak in West Africa. The Red Cross uses crowdsourced maps to deliver assistance to small or difficult-to-find communities in the affected areas, while Doctors Without Borders uses OSM to map the spread of the disease and coordinate field teams.
Crisis crowdsourcing may have its roots in the unruly hashtags of Twitter, but with each new disaster, organizations and data specialists are finding new ways to shape the chaos of the crowd into something useful.