Citizen Tech: Social Media in Disaster Response

by Amy Sample Ward on February 19, 2011

Post image for Citizen Tech: Social Media in Disaster Response

Today I’m in San Antonio, Texas, for the American College of Preventive Medicine annual conference. I’m joining a panel to talk about the response for support after the Haiti earthquakes last year. My contribution to the panel is to provide context about the use of social media in emergency and disaster response as well as an overview of some of the tools we saw deployed last year and we may see in the future.

Citizen-Powered Response

According to the American Red Cross,

A recent Red Cross survey asked 1,058 adults about their use of social media sites in emergency situations. It found that if they needed help and couldn’t reach 9-1-1, one in five would try to contact responders through a digital means such as e-mail, websites or social media. If web users knew of someone else who needed help, 44 percent would ask other people in their social network to contact authorities, 35 percent would post a request for help directly on a response agency’s Facebook page and 28 percent would send a direct Twitter message to responders.

Social media, like all technology, is developed by people. It evolves to meet our changing needs, to fit our changing lifestyles, and to integrate into the way we do our work. There are two types of media we will look at here: direct and indirect content.

Direct Content

The first example of direct content is the use of Wikipedia during the 7/7 bombings in London. Millions of editors on Wikipedia and it’s rise in public use was climbing. On the morning of July 7th, a page was created for the events taking place in central London and as both television news and personal witnesses revealed more information the page content grew – by the minute. (You can review the page edits here: http://en.wikipedia.org/w/index.php?title=7_July_2005_London_bombings&dir=prev&action=history ) Links to news stories were included as they emerged and the occasional personal edit of sentiments like “no one knows what’s going on” contributed to the live action growth of the page.

What’s so important or interesting about this? Wikipedia was a popular resource and established as a citizen-driven information source. The bombings created confusion and as people were able to put details together for themselves, they did so in a way that others could learn and follow the developments. Working together, citizens could piece together news stories as well as facts from witnesses as they emerged to create the most complete picture of events.

Another direct content example is that of the number of websites that emerged post-Hurricane Katrina.

Several websites were set up to help family members find out information about each other in the chaos. Some include the Red Cross, The Weather Channel, local newspapers, Craigslist, and others. Yahoo set up 100 Internet-linked computers at the Astrodome and developed a meta-search of evacuee registration websites. On September 11, despite having reunited several families, the National Center for Missing and Exploited Children had a list of 1,600 children listed as missing by their parents, or who were seeking their families.

Problems were that many survivors had no internet access, let alone electrical power, let alone computers or even computer literacy. There were also many sites so a searcher would have to go through several and sort through the many different search protocols and syntax. Another problem in theory is fraud, and another problem is that many sites only included last and first names which in a mass of several hundred thousand displaced persons obviously included many duplicates.

Indirect Content

Now for indirect content or Mash-ups. The first example is Ushahidi – originally designed as a tool for mapping reports of violence in Kenya after the post-election unrest in 2008. Reports of violence and of peace efforts could be placed via the web or mobile phone. It eventually garnered 45,000 users who took advantage of the easy-to-access tool to place reports. This proved to the Ushahidi team that their tool would be valuable to communities around the world. Since then, the platform (which is open source) has been modified for use in South Africa (mapping xenophobic violence), DR Congo, Vote Report India (to monitor the recent local elections) and more.

What’s important to note here is that unlike wiki or websites, the direct forms of content creation, with Ushahidi citizens only needed to know the number to which they send reports from their phone. The platform would then take their data, map it from their sending location, and create the other visual support data you see on the map.

Another example of indirect content is that of social networks. There are many different platforms, but facebook is by far the largest (now with over 500 million users and 700 billion minutes of use per month, and 70 translations). In 2009, for example, 2 trapped girls in Australia chose to post to facebook for help rather than dialing directly. A similar example is that of an Atlanta city councilman who chose to post a message to Twitter asking for a medic to respond to an unconscious woman on the street rather than to dial 911.

Why Social Media?

Whether it’s direct or indirect content, why do citizens turn to social media in moments of need or disaster? There are three core reasons:

Bandwidth: the simple fact that if we have a very limited amount of time, battery or other opportunity we can reach unlimited numbers of people that can help by posting to social media instead. That’s why the Atlanta councilman used twitter! His mobile phone battery was nearly dead so he chose to send a message to twitter and ask for help as well as for those that couldn’t help to spread the message, instead of calling 911 in the chance that he would end up on hold.

Response: studies show that people expect a response on social media. It is a social space where engaging and interaction is the constant action and so we believe, too, that if we were to need help and support that the community would take action quickly.

Power: lastly, and very importantly, we see our contributions making an impact. The photo or video taken on a phone and then posted to the web can change the conversation, alert news media to issues or new developments, and change the course of response. The same with live information, personal stories from witnesses, and so on.

Examples from Haiti

Let’s look at a couple of the main examples from the Haiti earthquake response. I will explain the use of mapping, fundraising, and crowdsourcing. Again, I’ve included these driving questions to help you frame the decision making that citizens employ both in asking for help and those indirectly impacted making decisions about which technologies they can use to help others.

Immediately after the earthquakes struck Haiti, the Ushahidi team adapted the platform for crisis reporting and mapping for the area. Anyone in or outside of Haiti could use the tool via the web or on a mobile phone to make reports with voice, text or video; the reports were then mapped and sources verified. Thousands of reports were placed via SMS. Those not in Haiti can also use the application as it aggregates news and actions to take.

If you want to see the data historically, be sure to visit the site and press “Play” on the map. It will play back all the data to show you where reports came from each day so you can see how the actions on the ground changed over time.

Most people are already familiar with the American Red Cross and the speed with which they jumped into action after the January earthquakes. Within just three hours of the earthquakes hitting Haiti on January 12th, the Red Cross had set up a mobile fundraising mechanism that let people donate $10 (or any other amount they wished) from their phone by sending HAITI to 90999. 100% of donations went directly to disaster relief. By the morning of the 14th, donations had already hit $3 Million.

The total number reached $32 million!

The ARC wasn’t only using sms though…

The Red Cross wasn’t just taking in mobile donations, it was leveraging the power of the web for information distribution, as well — between those following the news and the organizations, relief workers on the ground and the services they were delivering, and more.

On the ground, the Red Cross focused on “food, clean drinking water and other relief items such as hygiene kits, blankets, tarps, sleeping mats, tents and water containers. The relief items are helping more than 10,000 families (50,000 people) to date — with plans to increase this number. About 79 percent of the funds spent or committed by the American Red Cross have been for food and water; 18 percent have been for shelter; and the rest have been for health and family services.” They’ve used their website, blog, Twitter and Facebook, Youtube and more to keep telling the stories of those in need, those they’re serving, and how people watching the relief effort unfold can continue to contribute.

A final example is that of sissing persons support: The Extraordinaries launched a mobile application that lets users take advantage of moments of free time to volunteer via their phone.  At the time of the earthquake, the application had more then 50 organizations contributing volunteer opportunities to over 6,000 users (with 35,000+ micro-tasks already completed).

Once the earthquakes shook Haiti, the team at The Extraordinaries went into action, creating ways for people to turn a few minutes into incredibly important volunteering. The system had three components:

The Image Tagger — Volunteers sort through news photos coming out of Haiti and categorize (tag) them with keywords like “adult, child, alive, deceased.” Never before has there been a system that can bring together thousands of photos from across the web and have them sorted by live human beings (since no computer could know that there is a teenager in a photo).

The Matcher — They’ve engineered a system that matches faces of missing people to faces in news photos that we’ve sorted with the image tagger above. Volunteers look at a photo of a missing person, compare it to a news image, and see if they can find a match.

The Search Engine — As volunteers sort through images with the image tagger, they are fed into the Extras’ “search engine”. This system allows families to search through images taken post-earthquake in Haiti, and specify certain characteristics. For example, if a family is looking for their missing mother, they can use the search engine to find images that volunteers have tagged with “adult” and “female.” Their mother might be in one of those photos.

Thousands of volunteers donated time from their phones and computer screens to help reconnect families; 76,584 images were tagged, 8,137 news images collected, 746 possible matches found, and 24 matches close enough to contact families. It was a tremendous effort by the team and all the volunteers who donated time. The Extraordinaries are still analyzing their efforts and identifying ways to improve the system for future use, showing that every time you deploy your technology is a chance to learn and improve for the next time.

Last year, scientists revealed that it was in fact the chicken that came first, not the egg. And the same applies here: that it is the people that come first, not the technology. Regardless of how well-designed or technologically sound the tool, if people aren’t already using that tool then they won’t turn to it in a time of disaster. It is not just that people need to know of a technology, but they really need to have it already integrated into the way they communicate and share for adoption to take hold during a crisis.

It’s also important to note that just as it is the people that determine which tools are deployed, it is people who are at the core of the connections: people reaching out for help, and other people looking for tools to help those in need. As organizations, responders, and service providers prepare for disaster response and create strategies to continue improving, remember that we continue to increase the level to which the global community of citizens feels embedded and integral to local support. Volunteers stepped up to translate text messages, people who have never visited Haiti or knew anyone there started tagging photos to find missing persons. It’s incredibly important for organizations and responders to remember that just as those in need turn to the network for help, you can also turn to the network to help you respond.

Costs & Benefits

In a June 2010 survey of the DomPrep40, an advisory board of disaster response practitioners and opinion leaders, nine out of 10 respondents said they are not staffed to monitor social media applications and respond in a major event. Yet, 90 percent of respondents also felt that the public expects some action based on social media applications.

So, what are the costs and benefits of social media in emergency response?

At this point, social media is changing and growing as much as our environment both physically and politically – the options are so numerous they can be a real cost. Knowledge will continue to be an issue as well, whether it is knowledge of the issue or location or knowledge of what is really needed. As we saw with the Haiti response, language can be a huge barrier to response rate as we rely on people to help translate messages at an un-approachable rate. And finally, context; what is really happening on the ground and how to designate a hierarchy of response mechanisms and priorities (for example, the Red Cross had 79% of funds going to food and water alone.

But on the other hand, the sheer amount of data means we have far more information at our disposal to make more informed and more targeted decisions. We can tap into local knowledge because our reports are coming from the source. We also have real-time access both on the ground as well as around the world meaning information can develop at the rate to which it is surfaced – whether that is with an example like Wikipedia or in something like Twitter. And finally, measuring impact as it reflects back on the priorities for responders as well as those working to engage the rest of the world in targeted support.

What’s Next?

What we often think of as “social media” has given way to a larger movement, known as the Real-Time Web. We interact with our friends and colleagues in real time on and offline — either in the office or out at coffee, on Twitter or social networks. So, why shouldn’t our information, data, actions, and search happen in real time as well?  More and more, now we can.

What’s more, we don’t just expect to be able to learn what’s happening as it happens, like having news and updates about aftershocks and relief teams on the ground, but we also want to be able to take action in real time (and see the effects of our actions to help out). This is why tools that work across platforms and take advantage of mobile phone access have become the stars of real-time: we can donate instantly from our phone, we can help find missing people while riding the bus to work. Organizations involved can quickly unleash the power of the crowd to help them in disaster relief at the same time they’re providing food and water to those who have lost their homes.

You always get a bit of bad with the good, however, especially with the news and public attention of a disaster relief effort. The Haiti earthquakes were no exception. Scams and controversy emerged quickly, mainly because so many were taking advantage of the power of the real-time Web to get information and follow developments. Publicity, allegations, public statements and promises were all shared within the social media sphere – the examination process of Yele, the organization founded by Wyclef Jean, is a perfect example.

What the real-time Web has really changed is the way we are able to use the technology tools and systems we have in place, not our human processes. We have always felt compassion and an immediate call to action in times of need. Now we have technology catching up with our responses times.

I think we will see a continued growth of real-time action and information based services for emergency response and I’m hoping it means that organizations and responders can leverage that data and those tools to integrate them in the efforts they already take.

Slides

Here are my slides from the presentation. As always, you can also review the speaking notes by downloading the file for your own review.

Resources