Conference PaperPDF Available

Social Media Analysis in Crisis Situations: Can Social Media be a Reliable Information Source for Emergency Management Services?

Authors:

Abstract and Figures

Learning and understanding what happened before, during, and after a crisis is extremely important for the improvement of the response process. For this purpose, social media has become an important communication medium used by both the affected persons and the emergency management services (EMSs). However, in different crises, different information may be needed, and the information shared in social media varies in its usefulness: It could be highly critical or completely irrelevant to the rescue operation. Supplying the best possible up-to-date information is crucial to the EMS, whose actions based on that information may save lives and resources. This paper studies a particular use case of extreme weather in Norway and identifies the information needs, the problem faced by EMSs, and how they use social media. It, further, pinpoints what different social media analysis platforms can provide in this type of crisis. The results of the research are criteria that social media analysis should follow to address EMSs' concerns. The output of this work can be used to more precisely describe social media communication for crises and to design more efficient platforms for information retrieval from social media.
Content may be subject to copyright.
27TH INTERNATIONAL CONFERENCE ON INFORMATION SYSTEMS DEVELOPMENT (ISD2018 LUND, SWEDEN)
Social Media Analysis in Crisis Situations: Can Social Media be a
Reliable Information Source for Emergency Management Services?
Mehdi Ben Lazreg mehdi.ben.lazreg@uia.no
University of Agder
Grimstad, Norway
Narayan Ranjan Chakraborthy narayan.chakraborthy@uia.no
University of Agder
Grimstad, Norway
Stefan Stieglitz stefan.stieglitz@uni-due.de
University of Duisburg-Essen
Duisburg, Germany
Tobias Potthoff tobias.potthoff@uni-due.de
University of Duisburg-Essen
Duisburg, Germany
Björn Ross bjoern.ross@uni-due.de
University of Duisburg-Essen
Duisburg, Germany
Tim Majchrzak timam@uia.no
University of Agder
Grimstad, Norway
Abstract
Learning and understanding what happened before, during, and after a crisis is extremely
important for the improvement of the response process. For this purpose, social media has
become an important communication medium used by both the affected persons and the
emergency management services (EMSs). However, in different crises, different information
may be needed, and the information shared in social media varies in its usefulness: It could be
highly critical or completely irrelevant to the rescue operation. Supplying the best possible up-
to-date information is crucial to the EMS, whose actions based on that information may save
lives and resources. This paper studies a particular use case of extreme weather in Norway and
identifies the information needs, the problem faced by EMSs, and how they use social media.
It, further, pinpoints what different social media analysis platforms can provide in this type of
crisis. The results of the research are criteria that social media analysis should follow to address
EMSs' concerns. The output of this work can be used to more precisely describe social media
communication for crises and to design more efficient platforms for information retrieval from
social media.
Keywords: Social media, Crisis management, Emergency management services.
1. Introduction
Social media has become the de facto medium for public crisis communication [38]. It plays a
pivotal role in most crises today, from obtaining life signs from people affected to
communicating with EMSs [11]. Three types of information sharing on social media during a
crisis can be distinguished. First, from EMS to the public: EMSs relay situation updates,
BEN LAZREG ET AL. SOCIAL MEDIA ANALYSIS IN CRISIS SITUATIONS...
evacuation orders, possible dangers... to the public. Second, from the public to public: the public
uses social media to maintain contact with relatives, friends and loved ones, and to show support
for the community. Finally, from the public to the EMS: The public uses social media to report
problems, needs, calls for help, and provide information throughout the crisis. The first two
information sharing types are now well established with training given to EMSs on how to
effectively communicate information to the public [8][7]. Many social media platforms are also
providing ways for people in the affected area to report their safety status. The third type, on
the other hand, is still facing many challenges. Although the information shared by the public
can enhance the situational awareness of the EMS, many EMSs are still skeptical about using
that information.
A priori, the EMSs' skepticism might be due to the nature of the massages posted on social
media during a crisis: These messages tend to vary in their usefulness highly: The message can
be off-topic, personal, or informative. Much of the retrospective research on Twitter messages
posted during events show that those messages can identify the problems appearing during the
relief effort. Finding this useful information can accelerate disaster response. However, the task
is not easy due to, among other things, information overload. In an analysis performed on tweets
related to the 2015 Nepal earthquake, discovered that even though relevant topics are discussed,
the information present in the discussions is often irrelevant [24]. As an example, monetary
support is one of the most discussed topics, but the majority of the messages are appealing for
donations from ordinary people outside the affected areas and not actual financial needs which
are more relevant to the EMSs.
In addition, the way rumors and misinformation are spread on social media makes it an
untrustworthy source of information [31]. The social media analysis platforms can also be to
blame for the EMSs' skepticism: Most of the social media analysis platforms during a crisis
follow a data-driven approach that analyzes the data first by finding ways to extract as much
information related to the crisis as possible [31][32]. Such an approach may result in
information obtained during the analysis that is not useful to the EMS, as opposed to an
approach that inspects the data in ways that make it pertinent to the meaningful questions for
the EMS. Many of previously listed reason are founded. However, do they make social media
irrelevant as information sources for EMSs? In this paper, we contribute to bridging the gap
that exists in information sharing from the public the EMS in social media.
We will focus on an extreme weather use case in Norway. Around this use case, we conducted
an interview with governmental EMSs that include the police, firefighters, municipality, and
red cross. The aim of the interviews was to learn about the standard operating procedure in an
extreme weather crisis, how the EMSs share their information and establish situational
awareness, and what role social media plays in the emergency response process. Further, we
present an overview of the information social media can provide and the state-of-the-art social
media analysis platforms. By studying the data availability and needs from both the EMSs' side
and the social media side, we try to identify the common denominator between the two. This
paper contributes to addressing the question of how social media analysis platforms should be
designed to effectively support EMSs.
This paper is organized as follows. Section 2 presents our interview methodology and outcome.
Section 3 illustrates the diverse types of information social media can provide during a crisis.
Section 4 gives an overview of the state-of-the-art social media analysis platforms in crisis
situations. Section 5 discusses the lessons learned from our study and present the primary
outcomes of the paper. Finally, Section 6 concludes and provides pointers to future work.
2. Crisis response procedures and information needed
Extreme weather is the most damaging and frequent type of crisis in Norway (Norway
experiences on average 3 per year). It is characterized by an unusual and unexpected rainfall,
snowfall, heat or cold waves. When it occurs, extreme weather may damage the infrastructure
ISD2018 SWEDEN
including roads, leaving towns cut off from the rest of the country, the electricity network (for
example, during the storm Hilda in 2013, 35000 homes were out of electricity), and, most
importantly, it may lead to human injuries and death. The EMSs that are highly involved in this
type of crisis are the police, firefighters, municipalities and the Norwegian red cross. To find
out which information these EMSs need, we conducted a two-hour semi-structured personal
interview with representatives from four prominent local authorities on May 23, 2017: the chief
of staff at Agder police district, the crisis preparedness leader at Kristiansand municipality, head
of a unit at the Grimstad fire brigade, and a volunteer in Grimstad Red Cross. By using the
extreme weather use case, we managed to put the interviewees in a familiar situation in which
they have a lot of experience to get the most out of the interviews. The interview was oriented
to discover the current crisis response procedure, the information needed during this process,
the practice in information gathering, and the gaps of such practice. We brought up social media
during the interview to discover how the EMSs currently use it their opinions about its potential.
2.1. Methodology
Since we are concerned with the public-to-EMS information sharing through social media, the
interview aims first to understand the current crisis response and information gathering
procedure and identify the problems facing it. Further, we try to learn the extent to which social
media is currently used and what they think about its potential use in crisis situations. Therefore,
the questions we asked were oriented to achieve those previously cited goals. The questions
were:
How do you proceed during a crisis?
Which kind of information do you try to collect?
How do you obtain this information?
What are the biggest problems you face in this information collection process?
Do you use social media as an information source?
Would you be interested in analyzing social media?
For each of the above question, we asked a series of follow-up questions depending on the
participant's response. The analysis of the interviews was done qualitatively following the
Mayring approach [15]. When analyzing the interview data, we looked at the opinion held by
the EMSs on two critical question for this paper.
The first question is: How do they assess the current information gathering process during a
crisis? For this question, we distinguish three categories of opinions: high, middle, or low,
confidence in the process. Then, we classified the answers of each participant to one of the
previously listed categories. The second question is: What do they think about the use of social
media in crisis situations? Here, we separate between enthusiastic, halfhearted and skeptical
views. We further classified the interview data into one of these categories based on the opinion
each participant holds.
2.2. Findings
In Norway, during a crisis, police forces oversee the situation and responsible for the response.
The police operation center should have the overall picture of the crisis and act accordingly. In
the case of extreme weather, the most valuable information EMSs need to know are:
Which are the affected areas and people?
What are the areas in danger that need evacuation?
How fast is the water rising?
Who has already been evacuated and who still needs to be evacuated?
BEN LAZREG ET AL. SOCIAL MEDIA ANALYSIS IN CRISIS SITUATIONS...
Do the evacuated people have what they need?
Do individuals in a threatening situation know they need to evacuate?
Description of the current information gathering process
The information needed during an extreme weather crisis is diverse and sometimes very hard
to get promptly. To get that information, EMSs receive phone calls updated from all the
emergency call center, regional authorities, and other EMSs for updates about the situation: “In
current practice, we get our information through reports from emergency call centers. Besides,
we gather information by visiting some news, weather and media websites. But apart from that,
we do not collect information from other sources. The rest is information from the staff we send
to the event location,” the police representative said. The police officer agrees that there are
other techniques to gather information they can use to improve their assessment process.
“During search and rescue operations, the more useful information we have, the better the
operations. It is very crucial to gather as much information as possible as quickly as possible,”
asserted the Red Cross volunteer. The amount of information available early in the process that
helps direct the crews to the right spot, such as, for example, which places have been searched
for the last two hours, is crucial to the rescue operation. As it is now, the red cross representative
thinks that gathering the necessary information is a process that takes time: there is a lot of
information available in sources such as social media and news reports that can be more
efficiently collected using simple internet search that is not collected because the EMSs lack
the technological tools.
The lack of information during a crisis can lead to unexpected incidents for the EMSs. As an
example, the Kristiansand municipality representative shared with us his insight on a flood that
happened in Kvinsdal, Norway in 2014. The region experienced 130mm of rainfall in the
mountains over one night. To put this in perspective, 44.2mm is a normal rainfall in Norway.
The water took 12 to 18 hours to reach the town. When it did, the city bridge was damaged,
boathouses destroyed, and the cultural center submerged. The damages could have been
avoided with more coordination between the municipality, the firefighters and the police. Due
to this lack of coordination, no one had an overall picture of the situation and could assess the
extent of the crisis. This lack of a full picture and understanding of the situation caused the
authorities in place to allocate fewer resources than necessary. The civil protection, for example,
understood the extent of the water flood only when a civilian called to get help pumping water
out of his basement. The firefighters went from house to house trying to help the inhabitants,
but they had no idea about the overall extent of the flood. Having the right picture of the
situation allows the persons sitting at the top to make the correct decision based on the right
reasons,” the municipality representative asserted. He thinks that the main reason behind the
shortcoming in getting the overall picture is the lack of coordination and proactive information
sharing between different EMSs.
Potential use of social media in crisis situations
When we asked the interviewees about their experience with social media so far, they all agreed
that they only use social media to communicate updates to the public and monitor social
activities of other EMSs. However, when asked if it can solve some of the information needs
problems, their sentiments were mixed: The police representative was enthusiastic about the
idea. For him, it is important to include technology in crisis management work, especially
information from different data sources including social media. This information should enter
the control room, and the decision-makers should have access to those data. The Red Cross
representative was less enthusiastic, stating: If we can use social media or other information
sources to identify where our presence is needed most instead of sending persons on the ground,
that would help save us a lot of resources. But I cannot see how.This quote illustrates one of
the gaps that exist between EMSs and social media analysis: Despite the abundance of tool we
ISD2018 SWEDEN
describe in Section 4, the EMS officers we interviewed are either unaware of their existence or
do not see how these tools can be useful for them.
Finally, the municipality representative was more skeptical of the idea of using social media
for information gathering. He did not see how social media can help get a better overall picture
of the crisis. Despite the downfall that the current procedure might experience, he does not think
that social media is the answer. He stated thatwe only use information from different EMSs
to get an overall picture. We do not rely on `Mr. Somebody' for the information.” This statement
also reflects the trust issues toward the information shared on social media.
3. Information in social media
Data produced and shared in social media like Facebook, and Twitter has proven to be valuable
in many different contexts. The previous section reviewed the current information sharing
procedure for the EMSs we interviewed and the problems and gaps they present especially in
getting a comprehensive picture of the situation. In this section, we examine the various kinds
of information that are available and accessible on public social media platforms in specific
cases and argue that this information can help solve some of these problems.
3.1. Textual information
Though social media nowadays seem to give more weight to other types of information, purely
textual information is still at the heart of most platforms. Textual information is externalized,
explicit or codified and accessible with automated techniques [32]. Depending on the language,
textual information follows specific grammar rules which allow one to access even free and
unstructured text. Dictionary-based analyses allow the selection of relevant texts according to
a defined set of keywords or to assess sentiments in social media regarding a particular topic.
During the recent hurricanes Harvey and Irma in the United States and the Caribbean, a lot of
textual information was shared in social media. For example, the following Figure 1 depicts
information from the social networking site (SNS) Facebook about a person that needs to be
evacuated for special treatment due to her diabetes.
Fig. 1. Facebook post
In Twitter, people also shared information on where to find shelters (Figure 2). This behavior
has been observed in other crises, too [35]. This information can be useful for emergency
authorities as well, as they might not be aware of all shelters.
BEN LAZREG ET AL. SOCIAL MEDIA ANALYSIS IN CRISIS SITUATIONS...
Fig. 2. Twitter post
Often, the textual information is used to cluster the high number of posts into relevant and
irrelevant posts or further categories [37]. Textual information is often used in aggregation [16].
Still, the individual text may contain information that can be relevant for the response
management, e.g., about the inundation height in a particular area. The challenge for textual
information is to define and improve filters and other relevance criteria that limit the number
of posts to a manageable number.
3.2. Photos and videos
Most social media platforms are capable of sharing not only text but also pictures and videos.
Although the complexity of processing information in pictures or videos is much higher than
for texts, they can also inform emergency service agencies better. For example, Fohringer et al.
[5] use photos taken by eyewitnesses to derive quantitative data about water depth. Photos and
videos depict the real situation on-site. In contrast to textual information, multimedia data is
more objective and does not require laymen to interpret what s/he observes locally.
Figure 3 shows a photo taken from a small flood in Oslo, Norway, and published on Twitter in
August 2016. The picture illustrates to the emergency service agencies: The depth of water at
the particular place, a car stuck in the flood, the road in need to be closed, and that people do
not seem to be injured.
(Live) videos shared on social media platforms usually contain even more information than
photos [20]. Videos can better convey, e.g., weather conditions or crisis dynamics. In the case
of floods, videos could be used to measure the flow rate. New services like Periscope, which
offers an easy-to-use live broadcast, are expected to become ``game changers'' (p. 8) in the
disaster response management because intermediaries can be skipped and hence, transfer time
reduced [4].
ISD2018 SWEDEN
Fig. 3. Photo depicting a flood in Oslo from Twitter
3.3. Spatial information
Most social media platforms provide location data for shared information, especially if mobile
devices with built-in location sensors (e.g., GPS) are used. Spatial information can be used to
filter for posts sent from a particular area [37] or to visualize shared information, e.g., with a
map [6]. Often, geospatial information is not enabled by default, making the precise location of
information impossible. In such cases, textual information mentioning the city or street have to
serve as a proxy.
3.4. Response information
Comments, answers or commented retweets can contain information that complements the
original information. In Figure 4, the person first added @mentions to attract the attention of
authorities. Later, she wrote the full address.
Fig. 4. Added details in Twitter reaction
BEN LAZREG ET AL. SOCIAL MEDIA ANALYSIS IN CRISIS SITUATIONS...
Eventually, the reply feature was also used to report the successful rescue. Hence, response
information is a valuable information source as it contains information about the actuality or
allows a better assessment of the severity of the reported issue (see also next section).
3.5. Quantitative information
Although some of the information mentioned above is also quantitative, we specifically want
to point to other numbers available. Reactions to original posts, e.g., expressed as a Like on
Facebook, or a Retweet on Twitter, can contain information about the urgency of the textual or
other information. Hence, quantitative information could help to prioritize open issues reported
through social media [11]. The pure number of comments or answers further indicates
relevance. Lastly, before a crisis happens, emergency service agencies can use available
quantitative data to assess and improve their potential reach that is also based on followers and
like numbers of official accounts.
4. State of the art in social media use in crisis situations
Social media can provide near-real-time information for the emergency responders to make
effective decisions throughout various stages of the disaster management process [10][33]. It
also facilitates connectedness during the emergency and provides relevant and timely
information from both official and non-official sources [29]. This information act as an artifact
in the online environment [36]. After examining the different kind of information available in
social media in some specific cases, this section takes a more global view on the platforms
available for social media analysis and how it is used around the world.
4.1. Use of social media in crisis
Social media was used as a participatory media during Hurricane Katrina (2005); one of the
first natural disasters where social media use was noted [23]. People used online platforms to
help affected people by donating clothes, toys, etc., and social media helped emergency
responders to coordinate [30]. Palen and Liu [23] noted that social media also played a
significant role in finding the missing people in Hurricane Katrina. These cases marked the start
of social media uses in crisis situations by EMSs; after that, social media data, or more
precisely, Twitter data, has been heavily utilized in the management of almost every kind of
disaster [3], for example the 2009 Red River flood in the US [34], the 2010 Haiti earthquake
[27], the 2011 Japan tsunami [1], the 2015 Nepal earthquake [24] etc.
During the Haiti earthquake in 2010, digital volunteers used Twitter data to map the affected
areas using Ushahidi. This crisis map became an invaluable resource for relief workers in the
field [27]. During the 2015 Nepal earthquake, the Nepal police used social media as one of the
main communication channels [28]. Kathmandu Living Lab prepared a Nepal quake map based
on social media data [13]. The open nature of social media data helps the responders to operate
from the ground. Social media are now gaining attention among those dealing with extreme
weather disasters [19].
4.2. Social media analysis platforms
It is always challenging to analyze social media data because of the diverse sources and
unstructured nature of the data. Facebook, Twitter, and other social media platforms do not
provide similar kinds of data, thus the need for social media analytics tools to analyze this data.
There are few platforms available used by the volunteers to analyze this unstructured data of
social media during any crisis. The following list mentions a few standard tools.
Artificial Intelligence for Disaster Response (AIDR)
Tweak-the-Tweet
ISD2018 SWEDEN
Ushahidi
TweetTracker
TweetCred
The mentioned tools work for Twitter data sources or a combination of other social media
platforms. Artificial Intelligence for Disaster Response (AIDR) uses machine learning to
identify crisis-specific Twitter data automatically. By using a small of labeled tweets, AIDR
allows one to detect the categories of tweets [12]. Through the identification of the category of
tweets, it helps the responder to react to particular issues rather than the crisis as a whole.
Researchers at the University of Colorado developed a crowdsourcing platform called Tweak-
the-Tweet. It focuses on specific hashtags to make the data structured [14]. After that, a parsing
algorithm can be used to extract the information. One of the benefits of this system is that it can
work with the existing social media infrastructures [27].
Ushahidi is a crowdsourcing platform first developed to map the reports of Kenyan post-
election violence in 2008. Later it was widely used during the Haiti earthquake and Hurricane
Sandy. This system is not only used to collect Twitter data. It can also be used to gather data
from RSS feeds, email and SMS [26]. This variation makes this system more attractive.
To collect data from Facebook, Twitter, YouTube, etc. with a combination of keywords,
location, and user information, TweetTracker was developed. It can easily map the geotagged
post. There is a particular module in TweetTracker to facilitate disaster relief [18].
It is essential to get credible information during a crisis to assess the situation accurately. Due
to the dynamic nature of social media, fake news can spread quickly and create a mess on the
ground [21]. TweetCred was developed to find credible information shared by Twitter users in
real time. It provides a credibility rating of a post, and a supervised automated ranking algorithm
determined the credibility of that post [9].
5. Discussion and lesson learned
Social media is a platform with a significant potential to be used for public-to-EMS information
sharing. We showed in the previous section that the information available in social media during
a crisis can address EMSs' needs and help them establish a better situational awareness, and
complements the current information gathering procedures to get a complete picture of the
crisis. However, social media is still not considered a source of information for governmental
EMSs in Norway as well as in many other countries around the world. It is understandable that
EMSs question the value of social media for crisis response because of the gap still present in
social media analysis and research. In this section, we will pinpoint these gaps based on the
outcome of previous sections.
One of the main reasons behind the skepticism behind using social media as an information
source by EMSs is the lack of confidence in the information present in the platform: As
mentioned by our interviewees, they do not trustMr. Somebody” to deliver accurate situation
update and needs. This concern is founded on the amount of rumor and disinformation spread
in social media during a crisis [31], which leads us to the first question a social media analysis
platform needs to answer: can it establish a trusted network of people to get the information
from? A lot of research has been carried on developing an automated trust model for social
media networks based on user behavior and interaction with other [25] [17]. However, none of
these methods are integrated into the crisis-related social media platforms described in section
4.
Social media content varies in quality from excellent to spam and abuse. Once a trusted network
is established, we need to ask how we can ensure that only high-quality messages shared by the
network are treated by the platform? By quality of the message, we mean one that is clear,
readable, and concise. Information quality assurance models to identify high-quality
BEN LAZREG ET AL. SOCIAL MEDIA ANALYSIS IN CRISIS SITUATIONS...
information are another evolving topic of research in social media [2]. Nevertheless, their
integration in social media analysis platform for crisis response is still unsatisfactory.
Many data scientists are relatively new to the field of social media in crisis research. They are
knowledgeable about the management and analysis of large-volume data but lack the
understanding of the EMSs' needs. Data scientists tend to think that large volumes of social
media data alone will reveal patterns of behavior during a crisis. Moreover, the growth of
artificial intelligence and machine learning during the last few years has led to the emergence
of many artificial intelligence-based analysis tools [12][37][11]. These tools analyze the data
first by extracting as much information about as many topics related to the crisis as possible.
When the focus is on the data, and its volume, rigor in data collection becomes an afterthought.
In contrast, social media data must be analyzed in ways that provide relevant answers to the
question asked by the EMS, which usually triggers new data collection steps and questions. For
a social media analysis platform to be efficient in a crisis situation, it needs to focus on
answering EMSs' questions and needs. To summarize, the list of question that social media
analysis platform need to answer to be an efficient tool for EMS during a crisis situation are:
Can social media analysis platforms establish a trusted network of people to get the
information from?
Can social media analysis platforms ensure that only high-quality messages shared by
the network are treated by the platform?
Can social media analysis platforms analyses social media data in ways that provide
relevant answers to the question asked by the EMS?
Table 1. Complice of socail media analysis platfoms to the critera deduced from this research
Socail media analysis
platform
Trusted
network
Quality of
infomation
ensurance
Important
EMSs question
answering
AIDR
×
×
×
Tweak-the-Tweet
×
×
×
Ushahidi
×
×
TweetTracker
×
×
×
TweetCred
×
×
Table 1 shows how the current social media analysis platform discussed in section 4.2 comply
to the criteria discussed in this section. Many of these approaches are just data driven
classification of social media message. The table shows the gap that still exists between what
the social media analysis platform ca provide and what the EMSs require.
6. Conclusion
The information available on social media during a crisis, from textual information to
information in images and videos, can improve the current information gathering procedures of
the EMSs, help them get a better overview of the situation, assess the crisis response effort as
well as discover what the people in the affect areas need. Despite this information, EMSs in
Norway are still reluctant to use it as an information source. In this paper, we investigated the
reasons behind this reluctance via interviews with four major EMSs in Norway. As a result of
this research, we were able to assess what the social media analysis platforms currently lack
and should provide to address the concerns of the EMSs. The criteria are a trusted network,
information quality assurance, and answering the meaningful question for the EMSs.
ISD2018 SWEDEN
Acknowledgements
We would like to thank our interviewees (the Agder police district chief of staff, the crisis
preparedness leader at Kristiansand municipality, the head of a unit at the Grimstad fire brigade,
and the Grimstad Red Cross volunteer) for their collaboration and patience in answering all our
questions.
References
1. Acar and Y. Muraki. Twitter for crisis communication: lessons learned from japan’s
tsunami disaster. International Journal of Web Based Communities , 7(3):392–402, 2011.
2. E. Agichtein, C. Castillo, D. Donato, A. Gionis, and G. Mishne. Finding high-quality
content in social media. In Proceedings of the International Conference on Web Search
and Web Data Mining - WSDM ’08, page 183, New York, New York, USA, 2008. ACM
Press.
3. C. A. Cassa, R. Chunara, K. Mandl, and J. S. Brownstein. Twitter as a sentinel in
emergency situations: lessons from the boston marathon explosions. PLoS currents , 5,
2013.
4. E. Fichet, J. Robinson, and K. Starbird. Eyes on the Ground: Emerging Practices in
Periscope Use during Crisis Events. Proceedings of the 13th International ISCRAM
Conference, (Hardy 2013), 2015.
5. J. Fohringer, D. Dransch, H. Kreibich, and K. Schroter. Social media as an information
source for rapid flood inundation mapping. Natural Hazards and Earth System Sciences,
15(12):2725–2738, dec 2015.
6. H. Gao, G. Barbier, and R. Goolsby. Harnessing the Crowdsourcing Power of Social
Media for Disaster Relief. IEEE Intelligent Systems, 26(3):10–14, may 2011.
7. V. S. M. W. Group et al. Next steps: Social media for emergency response. Homeland
security science and technology, 2012.
8. V. S. M. W. Group et al. Lessons learned: Social media and hurricane sandy. Homeland
security science and technology, pages 1–39, 2013.
9. Gupta, P. Kumaraguru, C. Castillo, and P. Meier. Tweetcred: A real-time web-based
system for assessing credibility of content on twitter. In Proc. 6th International Conference
on Social Informatics (SocInfo). Barcelona, Spain, 2014.
10. Q. Huang and G. Cervone. Usage of social media and cloud computing during natural
hazards. Cloud Computing in Ocean and Atmospheric Sciences, pages 297–324, 2016.
11. M. Imran, C. Castillo, F. Diaz, and S. Vieweg. Processing social media messages in mass
emergency: a survey. ACM Computing Surveys (CSUR), 47(4), 2015.
12. M. Imran, C. Castillo, J. Lucas, P. Meier, and S. Vieweg. Aidr: Artificial intelligence for
disaster response. In Proceedings of the companion publication of the 23rd international
conference on World wide web companion, 2014.
13. Lambert. Disaster data assemblages: Five perspectives on social media and communities
in response and recovery. In System Sciences (HICSS), 2016 49th Hawaii International
Conference on, pages 2237–2245. IEEE, 2016.
14. Y. Liu and R. Raveendranath. Social media analysis for crisis management: A brief
survey.
15. P. Mayring. Qualitative content analysis/philipp mayring. In Qualitative content analysis.
Forum: Qualitative Social Research, volume 1, 2000.
16. M. Mendoza, B. Poblete, and C. Castillo. Twitter under crisis: Can we trust what we RT?
In Proceedings of the First Workshop on Social Media Analytics - SOMA ’10, SOMA
’10, pages 71–79, New York, New York, USA, 2010. ACM Press.
17. Mohaisen, H. Tran, N. Hopper, and Y. Kim. Understanding social networks properties for
trustworthy computing. In Distributed Computing Systems Workshops (ICDCSW), 2011
31st International Conference on, pages 154–159. IEEE, 2011.
18. T. H. Nazer, G. Xue, Y. Ji, and H. Liu. Intelligent disaster response via social media
analysis a survey. ACM SIGKDD Explorations Newsletter, 19(1):46–59, 2017.
19. B. Nerlich and R. Jaspal. Images of extreme weather: Symbolising human responses to
climate change. Science as Culture, 23(2):253–276, 2014.
BEN LAZREG ET AL. SOCIAL MEDIA ANALYSIS IN CRISIS SITUATIONS...
20. T. Norri-Sederholm, M. Joensuu, and A.-M. Huhtinen. Ensuring Information Flow and
the Situation Picture in Public Safety Organisations At Situation Organisations At
Situation Centres. In Proceedings of the 16th European Conference on Cyber Warfare and
Security, pages 267–273, 2017.
21. O. Oh, M. Agrawal, and H. R. Rao. Information control and terrorism: Tracking the
mumbai terrorist attack through twitter. Information Systems Frontiers , 13(1):33–43,
2011.
22. L. Palen and K. M. Anderson. Crisis informatics A new data for extraordinary times.
Science, 353(6296):224–225, 2016.
23. L. Palen and S. B. Liu. Citizen communications in crisis: anticipating a future of ict-
supported public participation. In Proceedings of the SIGCHI conference on Human
factors in computing systems, pages 727736. ACM, 2007.
24. J. Radianti, S. R. Hiltz, and L. Labaka. An overview of public concerns during the
recovery period after a major earthquake: Nepal twitter analysis. In 2016 49th Hawaii
International Conference on System Sciences (HICSS), pages 136–145. IEEE, 2016.
25. M. K. Rahman and M. A. Adnan. Dynamic weight on static trust for trustworthy social
media networks. In Privacy, Security and Trust (PST), 2016 14th Annual Conference on,
pages 62–69. IEEE, 2016.
26. Skuse and T. Brimacombe. Social networking, social media and complex emergencies:
Issues paper, 2014.
27. K. Starbird. Digital volunteerism during disaster: Crowdsourcing information processing.
In Conference on human factors in computing systems, pages 7–12, 2011.
28. R. Subba and T. Bui. Online convergence behavior, social media communications and
crisis response: An empirical study of the 2015 nepal earthquake police twitter project. In
Proceedings of the 50th Hawaii International Conference on System Sciences , 2017.
29. M. Taylor, G. Wells, G. Howell, B. Raphael, et al. The role of social media as
psychological first aid as a support to community resilience building. Australian Journal
of Emergency Management, The , 27(1):20, 2012.
30. C. Torrey, M. Burke, M. Lee, A. Dey, S. Fussell, and S. Kiesler. Connected giving:
Ordinary people coordinating disaster relief on the internet. In System Sciences, 2007.
HICSS 2007. 40th Annual Hawaii International Conference on , pages 179a–179a. IEEE,
2007.
31. R. M. Tripathy, A. Bagchi, and S. Mehta. A study of rumor control strategies on social
networks. In Proceedings of the 19th ACM international conference on Information and
knowledge management , pages 18171820. ACM, 2010.
32. Tuomi. Data is more than knowledge: Implications of the reversed knowledge hierarchy
for knowledge management and organizational memory. Journal of Management
Information Systems , 16(3):103–117, 1999.
33. D. Velev and P. Zlateva. Use of social media in natural disaster management. Intl. Proc.
of Economic Development and Research , 39:41–45, 2012.
34. S. Vieweg, A. L. Hughes, K. Starbird, and L. Palen. Microblogging during two natural
hazards events: what twitter may contribute to situational awareness. In Proceedings of
the SIGCHI conference on human factors in computing systems , pages 1079–1088.
ACM, 2010.
35. M. Yadav and Z. Rahman. The social role of social media: the case of Chennai rains-
2015. Social Network Analysis and Mining , 6(1):1–12, 2016.
36. D. Yates and S. Paquette. Emergency knowledge management and social media
technologies: A case study of the 2010 haitian earthquake. International journal of
information management , 31(1):6–13, 2011.
37. J. Yin, A. Lampert, M. Cameron, B. Robinson, and R. Power. Using social media to
enhance emergency situation awareness. IEEE Intelligent Systems , 27(6), 2012.
38. Zielinski, S. E. Middleton, L. N. Tokarchuk, and X. Wang. Social media text mining and
networkanalysis for decision support in natural crisis management. Proc. ISCRAM.
Baden-Baden, Germany , 2013.
Article
This paper presents a vision for a Disaster City Digital Twin paradigm that can: (i) enable interdisciplinary convergence in the field of crisis informatics and information and communication technology (ICT) in disaster management; (ii) integrate artificial intelligence (AI) algorithms and approaches to improve situation assessment, decision making, and coordination among various stakeholders; and (iii) enable increased visibility into network dynamics of complex disaster management and humanitarian actions. The number of humanitarian relief actions is growing due to the increased frequency of natural and man-made crises. Various streams of research across different disciplines have focused on ICT and AI solutions for enhancing disaster management processes. However, most of the existing research is fragmented without a common vision towards a converging paradigm. Recognizing this, this paper presents the Disaster City Digital Twin as a unifying paradigm. The four main components of the proposed Digital Twin paradigm include: multi-data sensing for data collection, data integration and analytics, multi-actor game-theoretic decision making, and dynamic network analysis. For each component, the current state of the art related to AI methods and approaches is examined and gaps are identified.
Article
Full-text available
In knowledge management literature it is often pointed out that it is important to distinguish between data, information and knowledge. The generally accepted view sees data as simple facts that become information as data is combined into meaningful structures, which subsequently become knowledge as meaningful information is put into a context and when it can be used to make predictions. This view sees data as a prerequisite for information, and information as a prerequisite for knowledge. In this paper, I will explore the conceptual hierarchy of data, information and knowledge, showing that data emerges only after we have information, and that information emerges only after we already have knowledge. The reversed hierarchy of knowledge is shown to lead to a different approach in developing information systems that support knowledge management and organizational memory. It is also argued that this difference may have major implications for organizational flexibility and renewal.
Article
Full-text available
The success of a disaster relief and response process is largely dependent on timely and accurate information regarding the status of the disaster, the surrounding environment, and the a ected people. This information is primarily provided by rst responders on-site and can be enhanced by the firsthand reports posted in real-time on social media. Many tools and methods have been developed to automate disaster relief by extracting, analyzing, and visualizing actionable information from social media. However, these methods are not well integrated in the relief and response processes and the relation between the two requires exposition for further advancement. In this survey, we review the new frontier of intelligent disaster relief and response using social media, show stages of disasters which are reflected on social media, establish a connection between proposed methods based on social media and relief efforts by rst responders, and outline pressing challenges and future research directions.
Conference Paper
Full-text available
This study aims to provide information about current, and possibly near-future communication interfaces in Public Safety Organisations' situation centres. Public Safety Organisations like the police, rescue and emergency services, as well as emergency response centres require a reliable and up-to-date situation picture at all times. In the event of demanding multiauthority operations, like crises, multi-casualty incidents, and emergencies, the role of Public Safety Organisations' 24/7 situation centres becomes vital. Typically, the different authorities have their own situation centres, which cooperate by using different technological solutions like a multi-authority radio network and real-time video connections. The use of new technology, information systems, UAVs, and body cameras facilitates receiving and processing information to build a more accurate situation picture. However, all the technology may cause moments of information overload. Furthermore, the increasing role of citizens in producing valuable information for Public Safety Organisations via social media, for instance, should not be overlooked. The role of information management is crucial in forming situational awareness and a common operational picture, especially in time-critical and rapidly evolving multi-authority missions. The literature related to Public Safety Organisations' situation centres is scarce, especially when explored from the information flow and situation picture point of view. In this paper, we explore the current technologies, practices, and action models in use to ensure adequate information-sharing between Public Safety Organisations to enable a common operational picture and collaborative decision-making in hybrid environments. The empirical data were collected through interviews with staff at different Public Safety Organisations and analysed using qualitative content analysis. The results suggest that the critical factors in ensuring information flow and thus the situation picture are communication tools, access to all the necessary channels and groups, common operational picture systems to enable shared situational awareness, and jointly approved cooperation plans combined with regular training and drills. In addition, human factors play an important role in complex, multi-authority missions. The possibility to deliver pictures and live video from the scene was highlighted and the data security issues related to it were mentioned as a challenge. It can also be concluded that in the near future the authorities need to develop tools and processes for utilising big data produced by citizens in social media.
Article
Full-text available
Social media has altered the way individuals communicate in present scenario. Individuals feel more connected on Facebook and Twitter with greater communication freedom to chat, share pictures, and videos. Hence, social media is widely employed by various companies to promote their product and services and establish better customer relationships. Owing to the increasing popularity of these social media platforms, their usage is also expanding significantly. Various studies have discussed the importance of social media in the corporate world for effective marketing communication, customer relationships, and firm performance, but no studies have focused on the social role of social media, i.e., in disaster resilience in India. Various academicians and practitioners have advocated the importance and use of social media in disaster resilience. This article focuses on the role that social media can play during the time of natural disasters, with the help of the recent case of Chennai floods in India. This study provides a better understanding about the role social media can play in natural disaster resilience in Indian context.
Conference Paper
Full-text available
Emergency response organizations such as the police are using social media as an additional channel to communicate with the general public in times of crisis. However, the utilization of social media with a free-style communication protocol by emergency response authorities for crisis response is still insufficiently understood. Who are the people they communicate with? What types of crisis-related information are shared with them? How do people react to these messages? How does the general public react to a social media run by a police organization? This paper uses the typology of convergence behaviors in emergency response as an attempt to categorize the public interaction with social media platforms. Furthermore, it uses the Situation Crisis Communication theory (SCCT) to analyze the crisis communication practices by a law enforcement agency. A content analysis of Nepal Police Tweets from the aftermath of the 2015 Nepal earthquake was conducted. It sought to understand (i) whether people who Tweet exhibit the same convergence behaviors as reported in the literature, (ii) how online participants communicate among and between the different crisis convergence behaviors, and (iii) what would be the best communication practices to assist crisis response efforts. Data show that convergence behaviors differed from intensity of appearances and communications and Tweets helped reconcile the differences in perceived emergency response needs between netizens and the authorities.
Article
Crisis informatics is a multidisciplinary field combining computing and social science knowledge of disasters; its central tenet is that people use personal information and communication technology to respond to disaster in creative ways to cope with uncertainty. We study and develop computational support for collection and sociobehavioral analysis of online participation (i.e., tweets and Facebook posts) to address challenges in disaster warning, response, and recovery. Because such data are rarely tidy, we offer lessons—learned the hard way, as we have made every mistake described below—with respect to the opportunities and limitations of social media research on crisis events.
Chapter
Social media data have emerged as new sources for detecting and monitoring disaster events. Several recent studies have suggested that social media data streams can be used to mine actionable data for emergency response and relief operations. Such massive and rapidly changing data present several new grand challenges to archive and extract critical validated information for various disaster management activities. The volume, velocity, and variety of the data require advanced computing infrastructure technologies for big data management. Cloud computing is proposed as the ideal platform to address these computing challenges. This chapter discusses the opportunities and challenges associated with using social media to gain situational awareness during disasters and the feasibility of using cloud computing to build a resilient and real-time disaster management system. Several real case studies and applications are presented to illustrate how both social media and cloud-computing solutions are used for various natural hazards.
Conference Paper
In responding to disasters, Twitter is extensively used, both for information exchange and mapping the crisis, among citizens, and in relation to national and international humanitarian responders. This paper reports Twitter analysis aimed at identifying the most pressing issues that arose in the short term recovery phase starting about a week after the Nepal earthquake, including the heretofore neglected topic of mismatch between international relief and local cultures. Based on Twitter data collected between April 30th and May 6th 2015. 1,074,864 raw messages apparently related to the Nepal earthquake were retrieved, filtered and analyzed. This exploratory adapts established frameworks for use in the dataset. The results show that our framework can identify several unique problems, including that disaster relief efforts can trigger negative sentiments when they are conducted without understanding of the local cultures.