Tag Archives: crowdsourcing

Crowdsourcing, Digital Volunteers, and Policy: New Workshop Summary from the Wilson Center

Post by: Kim Stephens

English: Woodrow Wilson International Center f...

English: Woodrow Wilson International Center for Scholars Español: Woodrow Wilson International Center for Scholars (Photo credit: Wikipedia)

A year ago this month the Commons Lab, part of the Wilson Center’s  Science & Technology Innovation Program, hosted a workshop with the goal of  “bringing together emergency responders, crisis mappers, researchers, and software programmers to discuss issues surrounding the adoption of… new technologies.”  The discussions included an in-depth review of crowdsourcing, specifically the use–as well as the reluctance, to use digital technology teams to aid in both message dissemination as well as data aggregation. The 148 page report from that meeting was released yesterday and is titled:  “Use of Mass Collaboration in Disaster Management” with a  focus on “opportunities and challenges posed by social media and other collaborative technologies.”

The Executive Summary states:

Factors obstructing the adoption of crowdsourcing, social media, and digital volunteerism approaches often include uncertainty about accuracy, fear of liability, inability to translate research into operational decision-making, and policy limitations on gathering and managing data. Prior to the workshop, many in the formal response community assumed that such obstructions are insurmountable and, therefore, that the approaches could not be adopted by the response community. However, it became clear during the workshop that these approaches are already being integrated into disaster response strategies at various scales. From federal agencies to local emergency managers, officials have begun exploring the potential of the technologies available. Stories of success and failure were common, but out of both came policy, research, and technological implications. Panelists shared strategies to overcome barriers where it is appropriate, but resisted change in areas where policy barriers serve a meaningful purpose in the new technological environment.

…Workshop participants identified the following activities as some of the more urgent research priorities:

  • Creating durable workflows to connect the information needs of on-the-ground responders, local and federal government decision-makers, and researchers, allowing each group to benefit from collaboration;
  • Developing methods and processes to quickly validate and verify crowdsourced data;
  • Establishing best practices for integrating crowdsourced and citizen-generated data with authoritative datasets, while also streamlining this integration;
  • Deciding on the criteria for “good” policies and determining which policies need to be adapted or established, in addition to developing ways for agencies to anticipate rapid technological change;
  • Determining where government agencies can effectively leverage social networking, crowdsourcing, and other innovations to augment existing information or intelligence and improve decision-making (and determining where it is not appropriate).

Crisis Mapping, Crisis Crowdsouring and Southern Storms

Post by: Kim Stephens

Photo courtesy FEMA photo library: Heckleburg, AL

A couple of weeks ago the #SMEMchat group discussed crowdsourcing and crisismapping and I’d like to revisit that topic again today. Whenever I give talks on this subject a lot of people indicate that they have never heard of crowdsourcing. But the wikipedia definition is fairly straightforward: Crowdsourcing is the act of outsourcing tasks, traditionally performed by an employee or contractor, to an undefined, large group of people or community (a “crowd”), through an open call. Jeff Howe, one of the first authors to employ the term, established that the concept of crowdsourcing depends essentially on the fact that because it is an open call to an undefined group of people, it gathers those who are most fit to perform tasks, solve complex problems and contribute with the most relevant and fresh ideas.

I think this definition is best understood by using an example. If you have ever watched the local news and heard the station ask folks to send in pictures of an event (usually weather) then you are witnessing crowdsouring. There is an implicit incentive structure here: the local news channel gets to choose from hundreds of pictures and does not have to hire a photographer; those who contribute get to have their picture shown on television. (There are many books and articles written on this topic–see my bibiliography, I have an entire section on the topic.)

Essentially there are two types of crowdsourcing during a crisis:

1. The provision of intelligence/information.

During this past week’s Senate Committee on Homeland Security & Governmental Affairs hearing on the importance of social media in emergency management, the Administrator of FEMA, Craig Fugate, alluded to crowdsourcing by referring to citizens as sources of information. The “task” that is being outsourced is simply the task of providing information from the field. This information answers the first question after a crisis: What just happened?

Washington, DC, April 22, 2009 -- W. Craig Fug...

Image via Wikipedia

In Administrator Fugate’s written testimony, he states: “We value two-way communication not only because it allows us to send important disaster-related information to the people who need it, but also because it allows us to incorporate critical updates from the individuals who experience the on-the-ground reality of a disaster.” In other words, the technology now exist to crowdsource the inflow of critical data regarding the situation after a crisis. Just like the local news station asking people send in pictures, emergency managers can have access to information that allows them to understand the event from the perspective of those immediately impacted, just by monitoring YouTube. For example, there were hundreds of videos of the tornados in Alabama on Youtube, posted in realtime, giving anyone with a computer or a smart phone a great perspective on the amount of damage that likely occurred. http://www.youtube.com/watch?v=8HhzDs1B4YA&feature=related

2. Gathering/Sorting/Making Sense of Crisis Data

If people in emergency management are uncomfortable with data provided by the crowd via social media, then this second type of crowdsourcing is even more uncomfortable: Asking the “crowd” to help gather and sort this data. The emergency management community, however, should understand this concept because it has become a reality after every recent disaster.

There are many organizations whose stated missions are to help organize the crowd as well as the data: CrisisCommons, Org9, Humanity Road, Crisismappers: StandbyCrisis Task Force; Sahana, Tweak the Tweet, and Ushahidi–I’m sure I’m missing a few. The mission of CrisisCommons and the sub-group CrisisCamp–whose co-founder Heather Blanchard also testified before the above-mentioned congressional committee, can serve as an example of the type of support these groups provide after a crisis:

…to connect a global network of volunteers who use creative problem solving and open technologies to help people and communities in times and places of crisis. CrisisCampers are not only technical folks like coders, programmers, geospatial and visualization ninjas, but we are also filled to the brim with super creative and smart folks who can lead teams, manage projects, share information, search the internet, translate languages, know usability, can write a research paper and can help us edit wikis.

The recent tornadoes provide a great example of all of the above mentioned groups’ work. One of the task they are currently performing can be boiled down to one sentence: matching need with the desire to help. This, of course, is fundamental task after a crisis: all disaster plans have Volunteer and Donations Management Support Annexes because it is well understood that if not well organized and planned for, volunteers and donations can sometimes hinder instead of help the response and recovery efforts. Furthermore, volunteers that are turned away can become vocal about not being “allowed” to help, creating a political problem for the responding organizations as well.

One of the main tasks these new volunteer organizations perform is what Patrick Meier, co-founder of CrisisMappers, describes as crowdfeeding: providing information from the crowd for the crowd–skirting or bypassing the “official” response organizations altogether.

How is this done? New communications technologies, such as social media, allow people to broadcast their needs to anyone willing to listen. Above is an example of a “tweaked” or MT–modified tweet, by a Humanity Road volunteer. The stated need was posted by the Salvation Army, who tweeted that they are serving meals for volunteers (the hyperlink provided allows those interested in helping serve get more information). RVAREGal simply tweaked the information so that it could easily be read by a computer for inclusion in a database. She added hashtags for #need, #info and location #loc. As this database of information is built, it can then be upload into a visualization platform, such as Ushahidi. This has been done in support of the southern storms and you can view it at Alabama Recovery Information Map.

The Ushahidi map is a not only a visualization of the information, but also creates an entire ecosystem: links to original source, the date it was submitted (which is key since this information does expire); a description of the information and additional reports with similar data; as well as a form to submit needs or resources directly.

In conclusion, I always like to ask: what are we learning?

  • Non-governmental organizations and the volunteer technical community are working to gather data from the crowd after a crisis and put it into platforms easily used and understood by the general public–they are not waiting for permission by any government agency, and they are usually not registered as a “VOAD”.
  • The public freely shares information about their situation (I need help/I can provide help) on social platforms that can be seen by anyone in the world–not just local response officials.
  • Response organizations could turn to the volunteer technical community for help in sorting through large amounts of data after a crisis in order to process it into usable information.

As Patrice Cloutier stated on the chat, “We ask the public to be prepared, with social media plus mobile technology, they also want to participate….embrace it!”

Resources:

For more information on the origins and descriptions of crisis mapping see “What is Crisis Mapping? An Update on the Field and Looking Ahead” by Patrick Meier.

International Conference of Crisis Mappers, Four Themes

 

Intensity map for the 2010 Haiti earthquake

Image via Wikipedia

 

Post by: Kim Stephens

The International Conference of Crisis Mappers (Oct. 1-3) brought together many different groups of experts, with disciplines ranging from geospatial intelligence to international humanitarian relief, to focus in part on lessons learned from the response to Haiti. At the start of  the conference Patrick Meier, one of the event organizers, asked participants to reflect on how communication, cooperation and coordination can be improved for future responses to large-scale disasters. He reminded the audience, “Haiti is a clear outlier…we should be inspired by the response…but should not use it as our only model for moving forward.”

By examining tweets from the audience made during the morning presentations, I was able to tease out several themes that emerged:

1. Response organizations in Haiti mostly had to rely on low-tech solutions for mapping since the country had little technology infrastructure.

  • Paper maps were the norm for those working the response on the ground.
  • Information gaps existed, according to some, due to a lack of technology and spatial awareness of both the citizens in the affected country and the response community.
  • Information overload on the ground is what stops most collaboration/coordination from happening.
  • No matter how many digital maps are made, ultimately, decisions are made from sitreps and verbal agreements on the ground. Therefore, there are real challenges in incorporating crowdsourced information into established organizations and data flows.

2. Finding information about the affected area was sometimes easier to gather from those not in the impacted zone (e.g. , those in the Diaspora), although getting that information into the hands of responders in the field was a   challenge.

  • Discovering information about a particular place doesn’t always have to come from  survivors. Sometimes a “local” can be physically very far from the disaster location (a large number of ex-patriots wanted to help and had first hand knowledge of the impacted area).
  • Ushahidi’s best accomplishment could very well have been crowdsourcing volunteers.
  • Crowdsourcing also occurred in traditional organizations. The World Bank’s Galen Evans described how over 600 earthquake engineers were essentially crowdsourced to analyze aerial data.
    • Everyone wants to help in a disaster, but we should think about what experts can be utilized during an event before the event.
  • Was the information coming from SMS texts reliable? Christine Corbane from the Joint Research Centre explained how their research found geo-tagged, crowdsourced SMS text messages highly correlated to the spatial distribution of building damage intensity in Port-au-Prince.

3. The crisis mapping community can add value to the response community, but processes to do so need to be established.

  • The crisis mapping community can translate each affected person’s story during a disaster into actionable data so that crisis managers can act. (See great new blog post by Gisli Olafsson on this topic.)
  • But, creating a common language among cartographers, humanitarians and beneficiaries is tricky: we need to develop baseline cartographic literacy.
  • The UN’s OCHA representative described how the emerging technology community and the humanitarian community don’t speak the same language. The one thing we have in common is that we all want to help. Questions remain:
    • Who coordinates the crowds?
    • How does the tech community fit into the UN cluster groups?
    • Can the crowd be used for data processing and data cleaning?
    • Why didn’t these groups coordinate during the event? Everyone was overloaded, pre-event coordination needs to occur.
  • The keys to success will be shared standards, shared situational awareness, and shared goals.
  • Standard operating procedures should be put in place to help govern this information sharing.

4. Affected populations or nations can and should be empowered to help themselves.

  • Ushahidi representative noted how empowering the local community in Haiti was key, and the locals eventually took over information curation.
  • The Grassroots Mapping Network discussed how simple technology can be employed for data anaylsis:  anyone can use their kit to fly a kite or balloon with a camera attached and gather data without the need for a satellite connection.  This inexpensive solution can help communities do their own mapping.
  • A representative from Development Seed discussed how we should think about needs first and technology second because there are many places that have limited technology capabilities. He stated:”This is why we are building really tiny software” and introduced maps on a stick, or maps on USB drives loaded with spatial data for low resource settings. This can help people in the field, even those that don’t have internet connections.

One tweet sort of sums up the day: Does better data come from improved technology or more meaningful engagement of locals?”

For a complete summary of each speaker, see the blog of  Jillian C. York who transcribed the talks as they occurred.

For a more complete after action report on Crisis Information Management during the Haitian earthquake: see: “Haiti and Beyond” by the ICT for Peace Foundation, March 2010.

Did Crowdsourcing work for the BP Oil Spill? One local official says “no”.

Image representing Ushahidi as depicted in Cru...

Image via CrunchBase

Post by: Kim Stephens

Last week I found an article on another blog, Emergency Management , about Santa Rose county in Florida that used the off-the-shelf mobile software called”Xora” to track the oil spill. Xora’s publicity explains how it works:

“… the Emergency Management team members go out on water patrol boats to scout for oil spill product or boom problems. When product is found, the team member photographs them and then fills out a form on a mobile phone, noting the type of product. The photo is attached to the mobile form, and both are submitted from the phone directly back to the EOC for follow-up. Xora automatically captures the product’s GPS coordinates for documenting precise location, thus giving the EOC real time reporting and pictures. The EOC can then decide how and when to respond based upon the type of products found and what resources are needed to respond to the situation. The pictures could easily be forwarded to other operational command areas to dispatch the appropri- ate equipment.”

This reminded me of the deployment of Ushahidi software for the Oil Spill Crisis Map, which is a visual representation of reports of oil and its effects.  One of the key differences between the technologies is that instead of only responders reporting oil, citizens are allowed and encouraged to contribute: “This map visualizes reports of the effects of the BP oil spill submitted via text message, email, twitter and the web. Reports of oil sightings, affected animals, odors, health effects and human factor impacts made by the eyewitnesses and the media populate points on a this public, interactive, web based map. The information will be used to provide data about the impacts of the spill in real time as well as document the story of those that witness it.”

I was interested in whether or not Santa Rosa county  integrated any of the information found in the Oil Spill Crisis Map into their GIS system, or even if any of the data in the Crisis Map was being used to deploy emergency response personnel to verify or follow up on the information. I contacted a Daniel Hahn in their emergency management office, here is his reply:

“We are not integrating any of our information into ushahadi, a site I have recently become aware of and looked into. We were using Xora so as to be able to photograph and pinpoint potential oil threats to our inland waterways. It was also used to show the condition of deployed boom and confirm or refute the presence of response vessels. One problem I see with ushahadi in its current form is that it appears that  anyone can post anything, and as we soon learned during Deepwater Horizon, EVERYONE had oil in their backyards, or saw it in the bay, when in reality it was something else entirely (Sargassum, June grass etc…).

I too have been very interested in the use of social media as an emergency management tool, and the recent Red Cross study gave me hope that we as an EM community might keep up with technology and societal norms. I think controlled mediums where information is put out by EM is the best form of social media. In this way EM can have followers, and control over what is put on the site (i.e., the ability to delete incorrect of erroneous information). I do not see this ability with ushahadi. Ushahadi is good for what it was created for, which if I am correct is capturing and reporting human rights abuses. As an example I pulled up Santa Rosa on Ushahadi and saw where someone had posted oil washing up on Baldwin County beaches, when Baldwin County is in Alabama, not Florida. An organization called surfriders posted they were doing dispersant testing off shore. Who are they and what are their qualifications (rhetorical question)?” Daniel Hahn, Santa Rosa County, Florida.

This official clearly is interested in social media but doesn’t necessarily trust the public to provide valuable or reliable information. I think it would be interesting to see if any emergency response organizations along the Gulf Coast utilized this tool.  Hopefully, the map was able to provide a way the citizens of the Gulf to testify, which is what Ushahidi means, what they saw and the impacts that they felt. However, a more in depth analysis will be needed in order to determine whether or not the Oil Spill Crisis Map had any impact on the response effort.

How can University Emergency Managers use Social Media?

Kim Stephens was invited to talk to the Emergency Management for Higher Education Final Grantee Meeting,  sponsored by the Department of Education, about the application of social media to emergency management. A copy of her slides (39 pages) is included here or in the Presentations and Papers section of this blog. The talk included information on the following topics:

  • How are Universities currently using social media?
  • How do we engage students in emergency preparedness through social media?
  • What are some social media best practices?
  • What is crowdsourcing and how could it be utilized to create interactive applications or multi-media tools in order to increase students’ emergency preparedness?
  • How is social media used for crisis response? How can personnel in the response community use these applications to increase our situational awareness? (examples from Ushahidi’s deployment in Haiti are included)
  • What tools exists to collect, filter, verify & aggregate data?