How do people use social technologies to communicate and work together after disasters? And how can those technologies be improved to facilitate that communication?
Those were two of the core issues addressed by Kate Starbird, assistant professor in the Department of Human Centered Design & Engineering at the University of Washington, during her talk at the recent GeekWire Summit. Starbird discussed the emerging field of crisis informatics — the study of how information-communication technologies are used during crisis events.
Watch the full video below, and continue reading for an edited transcript.
Kate Starbird: A few folks have asked me what we do. I tell people, my short one is we teach engineers how to talk to people before they build things, which apparently is a good idea. My research looks at this intersection of computer science and social science and very specifically at the use of social media during disaster events.
I want to start by having you all look at this photograph and think about this woman in a very funny hat who’s got her phone out. This photograph was sent a few years ago, end of October 2012, on the east coast, right after Hurricane Sandy came in. This woman is looking at her phone and she’s maybe sending a text or a tweet or an Instagram photo about that tree-car interaction that’s happening behind her in the background.
In the wake of this event, there was a huge social media response. Millions of tweets, tens of photos per second on Instagram, other platforms that were popular at that time saw a lot of use, and I’ve been doing research in this space for a while, and this guy, he was either a blogger or a journalist, it’s sometimes hard for me to tell the difference, he contacted me and he said, “Hey, I want to talk to you about how Hurricane Sandy was the first social disaster. Can you help me with this?” I said, “Absolutely, I can help you.”
You cannot lead with the fact that Hurricane Sandy was the first social disaster. First of all, everything about disasters is and always has been social. I don’t think we would think about them as disasters, except for the fact that they disrupt our social lives, they affect people’s lives, they disrupt what we want to do, and they interrupt our normal social dynamics.
Disasters are inherently social, and ever since we’ve had social media and the platforms that came before social media that are like social media, people have been using these platforms during disasters in all sorts of creative and fantastic ways to try to share information and help themselves and help others.
I’ve been looking at this intersection for a few years of social computing during crisis events, and by social computing, I mean all of these tools and platforms that help us share information with each other, with our friends, with the remote crowd, and not just the tools and the platforms, I’m focused on the human behavior that those tools and platforms enable.
I look at them in a lot of different kinds of crisis events. I do look at the natural disasters, earthquakes, hurricanes. I look at other extreme weather events, like an inch of snow in Seattle, and I know what kind of disruption that can cause around here. Also, human-made events. At this intersection, there’s a lot of opportunity. A lot of things we can do that we couldn’t do before. People can share information with their friends, their families, their neighbors, the remote crowd, with emergency responders, with journalists, in real time about what’s happening to them on the ground. We’re all armed with these mobile phones now, and we can use these platforms that we have to share information.
This could help people make better decisions, it can enhance our situational awareness, help us understand what’s going on, in a perfect world, in a perfect information space, if we could get at the pieces of information that we needed. These tools can also be used by emergency responders to share information with their many publics in real time, like this evacuation notice that went out during the Boulder Floods.
There’s many other things, but the third thing that I look at is online volunteerism, how these platforms and tools facilitate people coming together to help themselves, to help other people in new ways. I want to talk to you today a lot about that.
I want to start out today with this photograph, and I want you to think back. This is where the US Airlines Flight, I think 2009, landing on the Hudson, and these are the first responders to that. These are ferry boats that were in the area that came as fast as they could and started helping people get off that plane.
Sociologists of disaster have known for a long time that after disaster event, people will converge onto the scene and, among other things, try to help. Our first responders are often not the professionals we think about. It’s often everyday people coming onto the scene to try to help out. More and more we’ve been seeing that convergence happen in social spaces and online spaces during disaster events. I’ve been looking at digital volunteerism for several years.
I want to take a little moment. Today is actually a really hard day. Yesterday was a really hard day to be a disaster researcher and I want to say that my heart is right now going out to the community of Umpqua and Roseburg at this time. I’ve been studying disaster events for a long time, well maybe not a long time, but it feels like a long time, and I’ve studied many events, and I want you to know that I get excited up here about what I’m talking about, but I know that these are events that affected people’s lives and they can be pretty heavy.
I’m going to talk today about three events, and I’m going to focus on the cool things about humanity that happen during events. I’m going to focus on the positives, how people come together to help each other.
I want to start out talking about the 2010 Haiti earthquake, which again, was a catastrophic event. Hundreds of thousands of people lost their lives, were displaced, there were tens of thousands injured and in the wake of that event, people had a very acute need for food, water, shelter. The infrastructure of Haiti was pretty vulnerable to start with and it was totally decimated.
For many weeks, people were in great need and I, like many other people, I had just started in this space studying disaster response online. I went online to try to help and actually, with the folks that I was working with at University of Colorado at the time, and we ended up finding a bunch of other folks who were trying to do the same thing, use social media tools to help. I want to focus on their stories, because they are much more interesting than mine.
I want to tell you about Melissa Elliott, and I’m going to translate this tweet. I think you all can read it, but this is a re-tweet that was sent about seven days, eight days after the Haiti earthquake, and SpyDoctorMedia says, “I am stunned. We have gotten supplies in, saved people from rubble, brought them doctors. We have the best team. We are volun-twitters.”
This tweet might make you wonder who are these people, who is Melissa Elliott, where was she, and how was she moving doctors around in Haiti? I want to tell you how that happened.
Melissa Elliott, @MelyMello, is a woman in Ontario, Canada, I think French speaking Canada, and she was affected by this event. The language in Haiti is Haitian-Creole, but the second language is French, and so there’s a connection with response and some other things between French-speaking Canada and Haiti, so there was a lot of people at the time that were going online and trying to help, and trying to help in other ways from that area.
Melissa Elliott went online and she wanted to do something. She cared about this place. She cared about these people, and so she started finding people that were from Haiti and who weren’t there, and actually found people that were there, and tried to gather up their cell phone numbers, so she could add cell phone minutes remotely to their phones, so they would be able to use their phones as the communication was going on.
She did this. She had lots of numbers. She started adding them, she started recruiting help. She had too many numbers. She couldn’t fill them all. She started recruiting help online. She started recruiting, trying to get donations online, and over time, after a while, she started meeting people. One of them was [Dayjek Mal 00:08:17]. This was a kid, he was 17, he was living in Florida at the time, going to school, but he had family in Haiti, and he connected with @MelyMello.
He said, “Can you add these minute for my family members?” He sent her lists of phone numbers, and over time, she developed this connection to him and all of his family members in Haiti. Later, as she’s seeing people need food over here, people have a doctor over here, he’s wondering how he could help, she starts to connect this information through these people whose cell phones she had. She starts calling them to say hey, I heard this, can you help? She starts becoming a remote operator, and for a few days, before the formal response got set up, she’s moving aid around in Haiti, and she wasn’t doing this alone.
There were other people. Maybe not on a wide scale, but they were actually having real impact, and there was other people that were helping her, and we did some interviews and one of them said, “I think that’s when I went on Twitter and started tweeting, then I discovered a whole bunch of people tweeting for Haiti, and started doing it myself and building up connection as much as I could in order to try to save some lives, if possible.”
These folks actually came together in really fascinating ways. I do a network diagram here. users are the spheres, and I connect them with a line, if they were re-tweeting or mentioning each other during that event. We have a bunch of Twitter data.
We found there was a pretty dense network. We selected users from some of the certain hashtags they were using, and among some of these aid-based hashtags, there was a dense network of people who were sharing information with each other, and throughout the course of this event.
We went back and interviewed them and we said what were you doing, why were you doing it, who were you working with, and we talked to about 20 folks in this network graph, and we asked them, eventually, how many people in this network did you know before January 12, 2010 when the event happened? Know online, offline, either one. Three of the 20 people knew one other person in the network. We gave them a list of all of the people they re-tweeted or talked to. They knew one other person in the network before the event.
After the event, they came together, they started coordinating their work, passing off one piece of information to another, saying hey, I’m in Australia, I’m going to bed, can someone in the US take over this request and call this person and see if they still need help, and this is really fascinating behavior.
Emergent organizations happen all the time during disaster events, but here we’re seeing emergent virtual organizations come together in networks like this. In repeat events over and over again, I’m seeing other networks form. Some of the same people come back, but there’s new centers of gravity, new people come in, but after every event, there’s these networks and on Twitter and elsewhere of people coming together to help out. Lesson learned here, the crowd is appropriating social media and online tools to converge digitally, to connect and collaborate, to solve problems during crisis events.
A second event I want to talk about is Hurricane Irene in 2011. Maybe there’s some journalists in here. I mention this because it was a really interesting case, where the role of a few journalists maybe changed or evolved from what we saw.
2011, Hurricane Irene was a dud event in New York City, but if you lived in upstate New York or Vermont, that storm actually got caught on top of those areas, and another storm came in at the same time. There ended up being catastrophic flooding. Whole towns were washed away, a lot of the road infrastructure, the bridges were washed away, people were trapped, people were in need of food and all sorts of different things for quite some time. Family members couldn’t connect with other family members. In the rural Catskill Mountains, the communication infrastructure isn’t what we think about. There’s no mobile phone connections. Everything’s land-line. Power went down, and so it was a very tough event.
If you were from the Catskill area, and I’ll explain this later, no matter who you were, you were probably getting your information through this, at some level, through this crowd source effort that was going on.
I was doing this research with a PhD student of mine. She collected all of this data. She was there at the time. She’s from the Catskill Mountains, and she collected all this data. Dharma Dailey, I wanted to mention her.
All right. What was happening is there was this live blog that was started. I don’t know if everyone’s familiar with a live blog, but it’s this up-to-date information center where you can … I think they used the Cover It Live service, where they can do real time information, and this site ended up having reports about people being trapped, reports about people missing their family members, trying to figure out where they were. People would ask a question, does anyone know about this area, does anyone know how to get from this area to that area, can we access that? Thousands of messages were sent on this platform, and it became an incredibly invaluable resource to the people that were there.
It was actually started by two journalists at this very tiny, local online newspaper, called the Watershed Post, and they recruited some other journalists from some other outlets and they put together this live blog. They allowed locals to post, and then they recruited people from the remote crowd to help mediate, moderate, to say this goes in and that doesn’t go in and everything else, because they didn’t have the capacity to do it.
They set up this really interesting crowd sourced collaboration, where you had information coming from one group and another group helping to process it. We talk about them as being crowd-sorcerers. We see them as orchestrating this effort in a really fascinating way.
Another thing we found in this that’s really important when I talk to emergency responders and it’s something we don’t always think about, is how interconnected the information space was. We actually saw information on this live blog. A lot of people put it there. They found it on Facebook, they took it to the live blog. Or they talked to someone on the phone and they entered that information to the live blog.
We had an incident where, not an incident, the situation was the radio broadcaster would read out information that they were seeing on the live blog. They were broadcasting that on radio. No matter where you were, information was circulating in and across these different platforms. It wasn’t circulating, excuse me. People were moving it. People were actively moving information from place to place to help to bridge gaps in the infrastructure and access. When we think about should I use Twitter, should I use Facebook, what should I do? You should have a good strategy, but you should also know that these information spaces are incredibly integrated and information is moving in and across them.
These are our two big findings. We know that the profession of journalism is changing. It’s got a lot of pressure. We know that people during crisis events will take on new roles and they will use their skill sets in new ways. In this case, we saw journalists evolving their role from I’m a reporter to I’m a crowd-sorcerer. I have these skill sets, I know how to use these platforms, I know how to recruit people, I have local expertise, and I’m going to help during this event in this very particular way.
The last event I want to talk about really shortly is Hurricane Sandy, and I want to do that because it offers some interesting comparison points to Seattle in this sense: Urban area that was densely populated. It’s highly connected to online spaces, people are reliant on social media, people are reliant on mobile phone connections.
We saw some really interesting behavior during that event. I want to focus on one little piece, which was people that are trying to find gas in New Jersey, and this is behavior we see a lot, but in this case, people were trying to find gas, and they developed a hashtag on Twitter. They said hey, everybody should use this hashtag, and they started promoting it, #NJGas.
What people would do would they say hey, I just got gas here, the line was this long, the price is this, they are price gouging, they’re not price gouging, or whatever, and they would post that information with #NJGas. If somebody wanted to find gas, all they have to do is go on Twitter, search for that hashtag, and they could see this organized information space.
If you had low bandwidth, you have low battery, you need power, you need gas, that might’ve been tough.
Here comes a volunteer account. Someone else who was outside the area was like that might be hard for you. I’m going to start curating that information, so people don’t have to follow the #NJGas the whole time. They just have to write me a message and I’ll go figure out where they can find gas.
We see this volunteer stuff happening in two levels, where the locals … This happens in both events. The locals are sharing and seeking real-time information about on-the-ground conditions and then the remote crowd is providing this assistance in different ways to help the locals get the information they need at the time they need it.
All right. Here’s, I think, why I got invited. I don’t know why I got invited today, but here’s why I think. I know that this, in the middle of the summer, if you lived in Seattle, you read this or you heard about this article and you got really scared or maybe you don’t live in Seattle, you’re thinking about moving to Seattle, you read this article, you said no, I don’t really want to go there after all. It’s talking about the really big one, and they’re talking about what scientists call the M9 Earthquake, the magnitude-9 earthquake, that is overdue.
I think I was asked to think about what would we do in Seattle. What’s going to happen in Seattle if this happens, and the thing is is we don’t know when it’s going to happen, so the technology could be very different at that time. All sorts of things could be different. We could … Holy moly. We could prepare in some way, and prevent some of the things that might happen, but let me think about what might happen in the next few years.
I got this photograph from a UW study, a simulated photograph from a UW study in 2009. I think this is the viaduct falling down. I’m going to speak to this a little bit about what might happen in this event.
From what I know and what I’ve been seeing, what’s going to happen is that people are going to try to survive, and they are going to take the tools they have on hand, and they’re going to try to work with them, they’re going to work with … People are very pro-social after disaster events. They’re going to work with each other. They’re going to help themselves, they’re going to help others. In some cases, those tools are going to be the platforms, the social media technologies, other technologies we don’t have yet. They’re going to be using maybe new technologies that people in this room are working on. They’re going to be appropriating these technologies and using them in ways we don’t even know yet to try to help themselves and help others.
They’re going to come together in networks. This network is going to be there. One of the first things I’m going to do if something happens is I’m going to start trying to connect to these people. I know they’re going to be there and they’ll be willing to help me. I’m going to try to connect to the remote crowd in any way that I can, because I know they’re going to be able to help me more than I’m going to be able to help myself.
Thinking about what we know about social media use during disaster events, I want to pose a few questions to this crowd and get you all thinking about ways that you could participate in some of this and one is how do we design for emergent collaborations? In some ways, Twitter and Facebook and these things … Twitter, the legend has it, it was designed a little bit with an emergency case in mind, which is why it’s so useful … One of the reasons why it’s so useful in that context, but these platforms are already being used for this, so as we’re developing other platforms, we can think what about the emergency case? What about Brazilians during disaster events?
One thing i really want to think about is how do we establish hyper-local social networks. Something happens on UW campus, sad to say, one of the first places I’m going to go is Yik Yak. Never go there on a regular day. It’ll ruin your faith in humanity, but it’s a local network, and only people in that area can be on it. You’re able to connect in a way without the noise of the crowd and actually see what’s going on, and I think it would be really useful.
I’m also wondering what’s the Yik Yak that we can use in the future that uses wireless technology that we can connect with each other while the communication infrastructure is down, and how do we collaborate? We can’t just develop it and say in a disaster you’re going to use it. You have to have technologies that people are already using for other reasons, they already know how to use, that they can easily take off the shelf and start using in new ways during disaster events.
This broad question is how can we build technologies, including the tools, the platforms, and the policies to support resilience during disaster?
With that, I want to thank you. Yes, I know some of you are the people that are out there doing this stuff. If anybody wants to work with me or wants to have some advice on how you can make your technologies something that’s super useful during disaster events, just let me know.
Thank you very much.