Tag Archives: research

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).

Social Media’s Use in Emergencies: Research from Victoria, Australia

Post by: Kim Stephens

We often lament in the emergency management community that there isn’t enough quantitative data regarding the use of social media in disasters. A new report from Australia is helping to fill that void. During the 2011 Victorian Floods social media was difficult to ignore. The Office of the Emergency Services Commissioner and the Victorian State Emergency Service therefore  commissioned Alliance Strategic Research to conduct an independent research project to explore and document social media’s use during that event. The research objectives:

  • document social media comments during the Victorian floods
  • analyse and ascertain the nature of these comments
  • establish flows of information and recommend approaches for future events

I haven’t had a chance to read through the entire report yet, but they have a great video describing the major findings. Two things that stood out to me: one, people start talking about recovery issues during the height of the crisis; two, the tone of people’s responses are more positive than negative.

Here is their summary of major findings:

  • the key behaviour documented was spreading information through social media channels, with the information generally helpful and positive in its nature;
  • regional areas of Victoria are active in social media;
  • different social media channels were used for different types of communications at different times;
  • social media volume increases with the population of the affected area, severity and duration of events;
  • Twitter was the most active medium, and was used heavily by media outlets;
  • evidence of a one-to-one communications model, with community members engaging with each other individually.

This link will take you to their homepage where you can download the full report.

Social Media’s use during crises is becoming commonplace: What are we learning?

Post by: Kim Stephens

Social Media’s use during disasters is now a common occurrence, every new event has its related #hashtag. But what are we learning from these events? My colleague Claire B. Rubin and I were discussing the overarching question “is the frequency of use bringing new innovations, progress, and improvements to crisis communications?” I guess the crux of the question, though, depends on progress for whom? I think certainly the public has seen all of the above–innovation, progress and improvements, however, information dissemination isn’t necessarily propagated from traditional sources.

I’d like to highlight two important thoughts regarding what we might be learning, and I’ll continue with this line of thinking for the next couple of days.

1. Are crisis communications improved by the use of social media or are social media the source of rumors and missinformation?

A new empirical study conducted by a research team at Yahoo addressed the question: Twitter Under Crisis: Can we trust what we RT?. The study focused on tweets after the Chilean earthquake with a focus on “the dissemination of false rumors and confirmed news”.

Our analysis shows that the propagation of tweets that correspond to rumors differs from tweets that spread news because rumors tend to be questioned more than news by the Twitter community. This result shows that it is posible to detect rumors by using aggregate analysis on tweets.

By analyzing tweets of 7 different items that were later confirmed as true against tweets of 7 different items that were later confirmed to be false, the authors of the study were able to understand “false rumor propagation”.  Overall, when an item was true, retweets (over 95%) affirmed the information, with a very low percentage of tweets denying–less than 1%. However, when an item was false, such as the tweet announcing the death of artist Ricardo Arjona, the number of retweets decreased considerably and the number of tweets denying the information increased markedly (over 50%). In other words, the community self-corrected the information.  The authors conclude:

This result suggests also a very promising research line: it could possibly detect rumors by using aggregate analysis on tweets.

The article also has a good bibliography for those interested in further reading.

2. Is progress occurring regarding the use of social media for crisis communications?

Another interesting presentation about the use of social media in disasters is called  “Community in Crisis: What Governments can learn from the Boulder Community’s usage of Social Media during the Boulder Fire”. The data were put together by Tery Spataro from Orange Insights. In this presentation, she demonstrates how people relied on social media, in particular twitter and Google maps with displays of aggregated data, to get information about everything from the status of the fire, to shelters, donations and opportunities to help aid victims.

Citizens not only filled the information void during the response phase, but are now using social media to help each other recover.

Of note:

  • 985,099 people were reached with tweets related to fire (#boulderfire)
  • A community map of the fire, created by a citizen using Google Maps,  included information regarding everything from fire areas, evacuation areas, emergency response info, and photos, and had over 1.8 million views
  • (Not in the slideshow, but I also found) A Facebook page dedicated to thanking firefighters called “Fourmile Heroes” with 1,857 fans. It was created by a group of Boulder citizens to help local firefighters who lost their homes.

In contrast, I found

  • The Boulder Office of Emergency Management, as of today,  has 674 friends on Facebook.
  • Boulder OEM has 464 followers on Twitter–they are following 0 people. Their last tweet was Sept. 17.
  • Boulder OEM did eventually launched a comprehensive website with fire-related information (so we should assume their efforts were concentrated there).

The author of the study concluded that the EM community should have had ready “a Twitter account, Map, Facebook, and include Mobile communications”.

I would argue that this slide-show demonstrates how progress is occurring. The progress, however, is maybe not what we expected. Maybe the progress is in how citizens are learning to help themselves. Boulder could become the ultimate demonstration of a resilient community.

Relevant Research – articles from ISCRAM 2010

Posted by Claire

The Proceedings are now available from the 7th International Conference on Information Systems for Crisis Response and Management (ISCRAM).  The conference theme was Defining Crisis Management 3.0 and featured topics such as: Collaboration and Social Networking; Geo-Information Support, and Humanitarian Challenges. These research papers are a significant contribution, in my view, in that may become the foundation of new efforts in the coming years.

The Proceedings (45 pp.)  include both abstracts and full text of papers delivered), May 2010. The  proceedings are available here ISCRAM2010 -Proceedings or can be downloaded from Scribd.com.