Track 8: Social Media Studies
Red Tape: Attitudes and Issues Related to Use of Social Media by U.S. County-Level Emergency Managers
Linda Plotnick, Starr Roxanne Hiltz, Jane A. Kushma, Andrea Tapia
Social media are ubiquitous in modern society. Among their uses are to provide real-time information during crisis. One might expect that emergency management agencies in the U.S. make use of social media extensively to disseminate and collect crisis information as that is where the information flows most freely and quickly; yet, these agencies are not fully exploiting the capabilities of social media. A survey of 241 U.S. emergency managers at the county level shows that only about half of these agencies use social media in any way as of 2014. Most do not have any formal policies to guide their use. Of those that do have formal policies, about one quarter actually forbid the use of social media. This study describes the barriers that impede use of social media by these emergency managers, and the ways in which they are currently used, and recommends steps to improve this use.
Amanda Lee Hughes, Apoorva Chauhan
This exploratory study examines how fire and police departments used online media during the 2012 Hurricane Sandy and how these media can be used to affect trust with members of the public during such an event. Using trust theory, we describe how online communications provide a means for emergency responders to appear trustworthy through online acts of ability, integrity, and benevolence. We conclude with implications and recommendations for emergency response practice and a trajectory of future work.
Adam Flizikowski, Marcin Przybyszewski, Anna Stachowicz, Tomasz Olejniczak, Rafał Renk
Information about location and geographical coordinates in particular, may be very important during a crisis event, especially for search and rescue operations – but currently geo-tagged tweets are extremely rare. Improved capabilities of capturing additional location from Twitter (up to 4 times improvement) are crucial for response efforts given a vast amount of messages exchanged during a crisis event. That is why authors have designed a tool (Text Analysis TWeet lOcator – TAT2) that relies on existing open source text analysis tools with additional services to provide additional hints about people location. Validation process, complementing experimentation and test results, included involvement of end-users (i.e. Public Protection and Disaster Relief services and citizens during a realistic crisis exercise showcase. In addition, the integration of TAT2 with external tools has also been validated.
Thomas Ludwig, Christian Reuter, Ralf Heukäufer, Volkmar Pipek
To be able to take efficient measures in crisis management, it is essential for emergency services to get as much details about an actual situation on-site as possible. Currently content from social media plays an important role since those platforms are used to spread crisis-relevant data within the population. Our contribution presents a concept which supports the situation assessment practices of emergency services by collaboratively evaluating and by analyzing citizen-generated content from social media using a multi-touch table. The concept was implemented based on a Microsoft PixelSense and evaluated with 14 participants. The results reveal the impact of subjectivity of the participants, their positioning around the table as well as the uniqueness of social media posts on the collaborative situation assessment with multi-touch tables.
Irina Temnikova, Carlos Castillo, Sarah Vieweg
We present the first release of EMTerms (Emergency Management Terms), the largest crisis-related terminological resource to date, containing over 7,000 terms used in Twitter to describe various crises. This resource can be used by practitioners to search for relevant messages in Twitter during crises, and by computer scientists to develop new automatic methods for crises in Twitter.
The terms have been collected from a seed set of terms manually annotated by a linguist and an emergency manager from tweets broadcast during 4 crisis events. A Conditional Random Fields (CRF) method was then applied to tweets from 35 crisis events, in order to expand the set of terms while overcoming the difficulty of getting more emergency managers’ annotations.
The terms are classified into 23 information-specific categories, by using a combination of expert annotations and crowdsourcing. This article presents the detailed terminology extraction methodology, as well as final results.
Raquel Gimenez, Leire Labaka, Jose Mari Sarriegi, Josune Hernantes
This research identifies from literature principles of successful Virtual Communities of Practice (VCoPs) and explains how they have been fulfilled in the development of a VCoP that aims at contributing to knowledge sharing on natural disasters. The developed VCoP involves 70 experts in dealing with natural disasters from different hierarchical levels, organizations and nationalities of Europe. The VCoP has been developed within a European project from the 7th framework program. During the project three workshops were arranged for the members of the VCoP to know each other and to develop a living document. The living document is a web based tool used by the VCoP to share documents and insights, and it helps VCoP members networking. This paper provides direction for developing a VCoP to exchange lessons learned reports among crisis managers and first responders, and it identifies barriers that hinder the use of the living document.
Holger Fritze, Christian Kray
In summer of 2014 the city of Münster experienced an urban flash flood not seen before with such intensity in Germany. This paper investigates the subsequent governmental and ad-hoc community response actions with a focus on the chronologies of Facebook and Twitter usage. Interviews identified drawbacks of coordinating volunteers in social media ecosystems. Possible solutions to overcome issues related to the interaction of community and official relief activities are identified.
No Misunderstandings During Earthquakes: Elaborating and Testing a Standardized Tweet Structure for Automatic Earthquake Detection Information
Francesca Comunello, Simone Mulargia, Piero Polidoro, Emanuele Casarotti, Valentino Lauciani
Social media have proven to be useful resources for spreading verified information during natural disasters. Nevertheless, little attention has hitherto been devoted to the peculiarities of constructing effective tweets (and tweet formats), or to common users’ comprehension of tweets conveying scientific information. In this paper, social scientists and seismologists collaborated in order to elaborate and test a standardized tweet structure to be used during earthquakes, expanding on the results of a quali-quantitative research project. The tweet format is specifically designed to launch an innovative information service by Istituto Nazionale di Geofisica e Vulcanologia (INGV): tweeting the automatic detection of earthquakes with a magnitude greater than 3. This paper illustrates the steps of the research process that led to elaborating a tweet format that will be used in the next few months by the official Twitter account @INGVterremoti.
Apoorva Chauhan, Amanda Lee Hughes
We report initial findings around the Facebook and Twitter adoption trends of 840 fire and police departments affected by Hurricane Sandy. The data show that adoption increased during the time period directly surrounding Hurricane Sandy. Despite this increase, the creation of new online accounts since that time has been declining and overall adoption rates seem to be stabilizing. Lastly, the data report Facebook to be significantly more popular than Twitter as a form of online communication for these fire and police departments.
Hongmin Li, Nicolais Guevara, Nic Herndon, Doina Caragea, Kishore Neppalli Cornelia Caragea, Anna Squicciarini, Andrea H. Tapia
Microblogging data such as Twitter data contains valuable information that has the potential to help improve the speed, quality, and efficiency of disaster response. Machine learning can help with this by prioritizing the tweets with respect to various classification criteria. However, supervised learning algorithms require labeled data to learn accurate classifiers. Unfortunately, for a new disaster, labeled tweets are not easily available, while they are usually available for previous disasters. Furthermore, unlabeled tweets from the current disaster are accumulating fast. We study the usefulness of labeled data from a prior source disaster, together with unlabeled data from the current target disaster to learn domain adaptation classifiers for the target. Experimental results suggest that, for some tasks, source data itself can be useful for classifying target data. However, for tasks specific to a particular disaster, domain adaptation approaches that use target unlabeled data in addition to source labeled data are superior.
THE EMSC TOOLS USED TO DETECT AND DIAGNOSE THE IMPACT OF GLOBAL EARTHQUAKES FROM DIRECT AND INDIRECT EYEWITNESSES’ CONTRIBUTIONS
Rémy Bossu, Robert Steed, Gilles Mazet-Roux, Caroline Etivant, Fréderic Roussel
This paper presents the strategy and operational tools developed and implemented at the Euro-Mediterranean Seismological Centre (EMSC) to detect and diagnose the impact of global earthquakes within minutes by combining « flashsourcing » (real time monitoring of website traffic) with social media monitoring and crowdsourcing.
This approach serves both the seismological community and the public and can contribute to improved earthquake response. It collects seismological observations, improves situation awareness from a few tens of seconds to a couple of hours after earthquake occurrence and is the basis of innovative targeted real time public information services.
We also show that graphical input methods can improve crowdsourcing tools both for the increasing use of mobile devices and to erase language barriers. Finally we show how social network harvesting could provide information on indirect earthquake effects such as triggered landslides and fires, which are difficult to predict and monitor through existing geophysical networks.
Fostering Community Resilience through Adaptive Learning in a Social Media Age: Municipal Twitter Use in New Jersey following Hurricane Sandy
Yang Zhang, William Drake, Yuhong Li, Christopher Zobel, Margaret Cowell
Adaptive learning capacity is a critical component of community resilience that describes the ability of a community to effectively gauge its vulnerability to the external environment and to make appropriate changes to its coping strategies. Traditionally, the relationship between government and community learning was framed within a deterministic paradigm. Learning outcomes were understood to result from the activities of central actors (i.e., government) and flow passively into the community. The emergence of social media is fundamentally changing the ways organizations and individuals collect and share information. Despite its growing acceptance, it remains to be determined how this shift in communication will ultimately affect community adaptive learning, and therefore, community resilience. This paper presents the initial results of a mixed-methods research effort that examined the use of Twitter in local municipalities from Monmouth County, NJ after Hurricane Sandy. Using a conceptual model of organizational learning, we examine the learning outcomes following the Hurricane Sandy experience.
Robert Power, Bella Robinson, John Colton, Mark Cameron
This paper presents a user configurable monitoring system to track in near-real-time tweets describing fire events. The system targets fire related words in a user defined region of interest published on Twitter which are further processed by a text classifier to determine if they describe a known fire event of interest. The system was motivated from a case study that examined a corpus of tweets posted during active bushfires. This demonstrated that useful information is available on Twitter about fire events from people who are in the vicinity.
We present an overview of the system describing how it is initially configured by a user to focus on specific fire events in Australia, the development of a text classifier to identify tweets of interest, especially those with accompanying photos, and the monitoring system that can track multiple events at once.
Ntalla Athanasia, Ponis T. Stavros
The research presented in this paper attempts an initial evaluation of Twitter as an instrument for emergency response in the context of a recent crisis event. The case of the 2013 disaster, when typhoon Haiyan hit Philippines is examined by analyzing nine consecutive days of Twitter messages and comparing them to the actual events. The results indicate that during disasters, Twitter users tend to post messages to enhance situation awareness and to motivate people to act. Furthermore, tweets were found reliable and provided valuable information content, supporting the argument that Twitter presents a very good potential to become a useful tool in situations where rapid emergency response is essential.
Antonin Segault, Federico Tajariol, Ioan Roxin
In the last decade, people have increasingly relied on social media platforms such as Twitter to share information on the response to a natural or a man-made disaster. This paper focuses on the aftermath of the Fukushima Daiichi nuclear disaster. Since the disaster, victims and volunteers have been sharing relevant information about radiation measurements by means of social media. The aim of this research is to explore the diffusion of information produced and shared by Twitter bots, to understand the degree of popularity of these sources and to check if these bots deliver original radiation measurements.
Robin Peters, João Porto de Albuquerque
The use of social media during disasters has received increasing attention in studies of the past few years. Existing research is mostly focused upon analyzing text-based messages from social media platforms such as Twitter, while image-based platforms have not been extensively addressed hitherto. However, pictures taken on-the-ground can offer reliable and valuable information for improving situation awareness and could be used as proxy indicators for relevance. To test this hypothesis, this work explores various social media platforms, including image- and text-based ones in the case of floods in Saxony 2013, Germany. Results show that there is a significant association between disaster-related messages containing images and their proximity to the disaster event. Hence, the existence of an image within a social media message can serve as an indicator for high probability of relevant content, and thus can be used for enhancing information extraction from social media towards improving situation awareness.
Imen Bizid, Patrice Boursier, Jacques Morcos, Sami Faiz
Content shared in microblogs during disasters is expressed in various formats and languages. This diversity makes the information retrieval process more complex and computationally infeasible in real time. To address this, we propose a classification model for the identification of prominent users who are sharing relevant and exclusive information during the disaster. Users who have shared at least one tweet about the disaster are modeled using three kinds of time-sensitive features, including topical, social and geographical features. Then, these users are classified into two classes using a linear Support Vector Machine (SVM) to evaluate them over the extracted features and identify the most prominent ones. The first results using the actual dataset, show that our model has a high accuracy by detecting most of the prominent users. Moreover, we demonstrate that all the proposed features used by our model are indispensable to achieve this high accuracy.
Gerhard Backfried, Christian Schmidt, Gerald Quirchmayr
Many possible links and connections can be observed between the different types of media used for communication during a crisis. These links can be detected and assembled to provide a more complete picture of events. They can be categorized according to the type of destination which yields important information for the gathering process as well as concerning general patterns of how platforms are connected. Tweets, posts and comments thus become parts of larger, linked sets of documents forming compound-documents. These documents stretch across media borders and platforms and provide context and broader information for individual entries. In the current paper we describe some of the links and linking behavior encountered during the floods in Central Europe of 2013 from the perspective of Twitter and Facebook.
Yang Ishigaki, Yoshinori Matsumoto, Yutaka Matsuno, Kenji Tanaka
We developed a series of inexpensive but accurate mobile radiation detectors, which we named Pocket Geiger (POKEGA), to address the urgent desire of ordinary people to measure and share radiation levels in their milieus and to discuss the results of the Nuclear Disaster in Fukushima, Japan. This action research reports on a new style of pragmatic model of radiation monitoring, which employs the features of Participatory Design and Participatory Sensing and adopts modern communication platforms such as crowd-funding, open source development, and Facebook. This paper proposes an interaction model between the project management body, and other inclusive corroborators, e.g., ordinary users and experts, and focuses on three development phases of the project: start-up phase, evaluation phase, and operation phase. This paper also considers a reliability assurance model on disaster information sharing between the citizen layer and the official layer by data sharing and discussion activities in the POKEGA community.