Big Data from Social Media and Crowdsourcing in Emergencies

2016
Anna Boye Koldaas
MSc in Security Risk
Management
University of Copenhagen
[BIG DATA FROM SOCIAL
MEDIA AND
CROWDSOURCING IN
EMERGENCIES]
Memo for the Danish Emergency Management Agency (DEMA)

2
Anna Boye Koldaas is an intern at the Communication Unit at the Danish Emergency Management
Agency (DEMA) Spring 2016. The internship is a part of her Master of Science in Security Risk
Management at University of Copenhagen. This memo is a part of the internship project exploring how
DEMA can use big data from social media and crowdsourcing to improve situational awareness in
emergencies. Read more about DEMA at http://brs.dk/eng/Pages/dema.aspx.
For comments or questions please email Anna Boye Koldaas at ako@brs.dk.

3
Content
Big data from social media and crowdsourcing in emergencies ............................................................... 4
Four ways of improving situational awareness .......................................................................................... 6
1. Automatic filters................................................................................................................................... 6
2. Digital volunteers................................................................................................................................. 7
3. Nudging and citizen-to-citizen emergency management.................................................................. 9
4. Crisis mapping .....................................................................................................................................11
How to exploit the full potential?.............................................................................................................. 13
Challenges and risks................................................................................................................................... 15
A critical note .............................................................................................................................................16
How can DEMA use big data from social media and crowdsourcing to improve situational awareness in
emergencies?...............................................................................................................................................18
Sources........................................................................................................................................................19
Figures......................................................................................................................................................... 21

4
Big data from social media and crowdsourcing in emergencies
Big data from social media and crowdsourcing can improve situational awareness in emergencies. The
Danish Emergency Management Agency (DEMA) defines an emergency as a situation that cannot be
handled by normal procedures and resources. Situational awareness is the perception and
understanding of elements in the environment and their likely developments in the near future.
Emergencies are often chaotic, involving the coordination of many agencies, and using data from the
public can improve situational awareness and assistance to people in need. Furthermore, social media
and crowdsourcing tools can become critical communication channels if conventional communication
breaks down because of overload or damaged infrastructure, such as after the attacks in Brussels in
March 2016.
Emergency management agencies and other agencies involved in emergencies mainly use social media
to inform the affected public and communicate with worried citizens. However, in more and more
emergencies, agencies use big data from social media and crowdsourcing to improve situational
awareness by monitoring trends, calls for help and plotting information on maps. This was done during
the violence following the election in Kenya in 2007, the Haiti earthquake in 2011, the hurricane Sandy
in the US in 2012 and the Nepal earthquake in 2015, to mention some. In all these cases, digital
volunteers filtered and sorted vast amounts of information, discovering trends and looking for calls for
assistance.
Emergency management agencies can use big data from social media and crowdsourcing to improve
situational awareness in emergencies in several ways. This memo discusses four ways of improving
situational awareness. First, automatic filters based on algorithms and the “normal” rate of posts on
social media can track trends and sudden spikes in activity indicating both the severity and
geographical extent of an emergency. Second, emergency management agencies can use digital
volunteers as gatekeepers to filter and sort information from social media that can be given to
emergency managers on the ground. Third, agencies can encourage people to participate in emergency
management and facilitate citizen-to-citizen emergency management by, for example, “nudging”
people to use certain trending hashtags, such as #shelter. Fourth, emergency management agencies can
create crisis maps by asking public to submit reports of the situation on the ground, such as fallen trees
during a storm, or use volunteers to go through big data from social media and plot the information on
a map. How can the emergency management agencies exploit the full potential of big data from social
media and crowdsourcing? How can agencies adapt to a future where emergency management
increasingly happens in a world with smartphones? Most emergency management agencies have
change the way they operate, so this change must be well informed and happen after carefully
considering both possibilities and risks.
Despite all the potential in using information provided by the public, emergency management agencies
also face some challenges and risks. The main challenge is to bridge the gap between the potential for
emergency management agencies to use digital volunteers and volunteers’ understanding of agencies’
needs. Another main challenge is the amount of information that is available, the majority of which is
noise and distraction. A third challenge is the accuracy and reliability of the data. How can emergency
management agencies know that the data reflects reality? The absence of reports does not necessarily
mean that there is nothing going on. Misinformation, rumours and trolling are other challenges to
utilising big data from social media and crowdsourcing during emergencies. Furthermore, if emergency
management agencies rely too much on social media, they risk that individuals use social media to ask

5
for help when they should be using conventional emergency channels. In addition to the challenges
and risks, researchers and practioneers debate the real effect of utilising big data from social media and
crowdsourcing in emergencies.
This memo is part of an internship project exploring the possibilities for DEMA to use big data from
social media and crowdsourcing to improve situational awareness in emergencies.

6
Four ways of improving situational awareness
1. Automatic filters
One way of using big data from social media in emergencies to improve situational awareness is to
create automatic filters that can track trends and sudden spikes in activity based on algorithms and the
“normal” rate of posts. A lot of the posts on social media are open to everyone, allowing emergency
management agencies to follow trends and developments. By registering the amount, frequency and
type of posts in certain locations and posts using certain hashtags, such as earthquake or storm, under
“normal” conditions the agency can create algorithms that can trigger alerts when an emergency causes
the activity to increase above the normal. Automatic filters can be combined with maps to show the
severity and geographical extent of an emergency. Automatic filters can give emergency management
agencies early warning signals, a better understanding of how an emergency develops and indications
of which areas an emergency is impacting. Obviously, automatic filters are only useful for populated
areas with relatively high internet penetration.
The U.S. Geological Survey (USGS) uses Twitter to monitor earthquakes. The USGS uses Twitter’s
Public API to create a baseline for earthquake tweets. USGS has found that people tweeting about
earthquakes make short tweets, and that first hand reports are unlikely to include links or numbers.
The USGS automatically filter the twitter stream based on these two criteria to determine when
earthquakes occur globally. Usually, the data from Twitter trigger an alert in less than two minutes
after an earthquake occurs. USGS can also use Twitter to detect false alarms from earthquake sensors
because if they’re getting reports of an earthquake in a populated area but no Tweets from there, that’s
a good indication that the sensor has triggered a false alarm.
Figure 1: The USGS using Twitter to track the earthquake in Mexico in 2011.

7
2. Digital volunteers
Emergencies generate a massive volume of big data from social media and emergency management
agencies do not have the capacity to process all this information. Automatic filters can improve
situational awareness, but computers and algorithms can miss important information and are unable
to the register the details of posts. As a response to the increasing amount of data from social media in
emergencies, groups and networks of digital volunteers has emerged across the world, such as Crisis
Mappers, Digital Humanitarian Network and Virtual Operations Support Teams (VOST). Emergency
management agencies and humanitarian organisations can request help from these networks in an
emergency or build their own networks of digital volunteers. Emergency management agencies can use
digital volunteers to filter and sort information from social media, the news and reports from the
public and generate information that is useful to emergency managers and response teams.
Emergency management agencies that want to use digital volunteers should consider what type of
volunteers they wish to engage with and in what way. People become digital volunteers for several
reasons, such as the wish to help, being personally affected by an emergency, the identity as a digital
volunteer, the digital volunteer community, personal relationships with other volunteers, and the
opportunity to learn new skills. Some only volunteer for a short time or during a specific emergency,
whereas others volunteer many times. In several emergencies, such as the Haiti earthquake in 2011, a
spontaneous group of volunteers formed and expanded as the needs grew. Many volunteers are
organised through networks. In some of these networks, anyone can become a member and the
members are requested to sign up for tasks when an agency or authorities request the help of the
network. In other networks, digital volunteers are organised into teams with a more hierarchical
structure and stricter entry requirements.
Using digital volunteers requires considerable effort from emergency management agencies. To ensure
effective cooperation and resources use, the agencies must provide a clear point of contact and have
staff available that can give feedback to volunteers quickly. Agencies should keep in mind that many of
the networks of digital volunteers operate around the clock, meaning that agencies must have staff that
is available at all times. Emergency management agencies that cooperate with international networks
should also develop structured training materials and templates that allow the networks to quickly
integrate new volunteers. In order to reduce the amount of required communication during an
emergency, the agencies have to clarify responsibilities, priorities and methods of work before an
emergency occurs. Consequently, agencies should develop strategies and procedures for using and
cooperating with digital volunteers. In most emergencies, volunteers have used various open access
tools to process information and communicate with each other. Emergency management agencies have
to consider how the information that the volunteers generate can feed into existing information and
emergency management systems.
The major challenge for agencies in using digital volunteers is the gap between the potential to use
digital volunteers and volunteers’ understanding of agencies’ needs. On one side, many emergency
management agencies have yet to realise and exploit the potential of digital volunteers and information
provided by the public. Using this resource requires adapting the organisation to the importance of
citizens in emergencies. One the other side, most networks of digital volunteers are spread out across
the world and do not have any previous knowledge about the specific emergency they are activated to
support or the agencies they are supporting. Consequently, national emergency management agencies
should consider whether it is effective to use groups from international networks of digital volunteers

8
or if they should develop their own groups of more closely selected and trained volunteers. Using big
data from social media involves a trade-off and the responsibility during an emergency is always with
the emergency management agency and not with the volunteers. The agency takes the risk that the
overall benefits of the system outweigh the time spent going through and responding to inaccurate or
faulty reports.
The American Red Cross (ARC) uses digital volunteers to improve situational awareness and
communicate with affected populations. The ARC has three Digital Operations Centers that they
activate in emergencies and use to communicate with volunteers and display the “digital information”.
Digital volunteers assist the ARC in gathering information from affected areas, spotting trends, and
connecting people with the resources they need, like food, water and shelter or giving emotional
support. In addition to notifying the ARC staff at the Digital Operations Center about trends and
developments, volunteers tag some posts for follow-up by ARC staff. The ARC also uses automatic
filters on data from social media to create word clouds and heat-maps that shows the activity in an
area. The word clouds are useful to discover places that are mentioned often and probably need
assistance and the heat maps can be useful to show the extent and centres of, for example, storms.
These visualisations based on the automatic filters are always updated, even when the Digital
Operations Centers are not activated. The ARC Digital Operation Centers shows how emergency
management agencies can use digital volunteers and tap into the potential of big data from social
media.
Figure 2: The American Red Cross Digital Operations Center.

9
3. Nudging and citizen-to-citizen emergency management
When an emergency strikes, most people turn still to conventional media and information channels,
but people also turn to social media for information and support. The Paris attacks in 2015 generated
over 10 million tweets using the hashtags “PrayForParis, #Paris and #ParisAttacks from November 12 to
151
. In countries were most people have smart phones or frequent access to the internet, the activity on
social media during emergencies will only increase. An important part of adapting to the increasing
activity on social media is to learn how to influence and shape the stream of posts. Emergency
management agencies can take on a “puppet master” role and attempt to “nudge” people to both to
generate useful information and information that is easy to track, but also to facilitate citizen-to-
citizen emergency management. Agencies can use social media to change the behavioural norms in
emergencies to create citizens that actively participate in emergency management.
Agencies can “nudge” people in several ways in order to influence the stream of posts on social media.
“Nudging” means to steer people’s actions in certain directions without limiting their choices2
or
without making some choices more costly in terms of time, money, trouble, social sanctions and so on3
.
One way to “nudge” people is to run campaigns teaching people about using hashtags in emergencies.
Agencies can create hashtags for different types of emergencies that are on their webpage and that
relevant agencies and volunteers can promote during emergencies. The Filipino Government has had
an official strategy on promoting the use of crisis hashtags since 2012. During periods of monsoon rain
and typhoons, the government has used #rescurePH, #reliefPH and other location specific hashtags to
improve situational awareness and response, for example to create lists of those in need of rescue.
Agencies can also develop guides to social media for citizens containing do’s and don’t’s in
emergencies. A second way to “nudge” people is to develop an emergency webpage or insert social
media streams on the emergency management agency’s webpage based on certain emergency hashtags,
so people can see their posts and follow similar posts. Seeing their posts can motivate people to
contribute and allows them to easily see what others post about the emergency. A third way to “nudge”
people is to pick up on popular hashtags already in use and encourage people to use these hashtags,
and in this way promote hashtags that are intuitive for people to use. Successful nudges are often
messages like: “Many people use #wildfires to document the emergency. How is the situation is where
you are?”. Emergency management agencies that have apps or warning systems can also use these to
encourage people to post reports or photos about the situation on the ground.
Emergency management agencies can influence the posts on social media to improve the usefulness of
the information posted, but to facilitate citizen-to-citizen emergency management. Through their
channels on social media, emergency management agencies can “set the standard” by encouraging
people to use specific hashtags, such as #needshelter, to help each other during emergencies. Agencies
can also promote local emergency management initiatives, such as food and shelter at a local
community centre. During the Dresden floods in 2013, a Facebook page promoting and facilitating
citizen-to-citizen flood relief quickly got over 40 000 likes. During the Boston bombings in 2013, people
offered food and shelter to people that could not get home through documents in Google Docs. Often
neighbours and locals can offer assistance quickly and can solve the situation without the aid of official
agencies. Citizen-to-citizen emergency management initiatives can utilise important resources and
provide relief efforts when and where emergency management agencies do not have the capacity to do
1
http://www.pri.org/stories/2015-11-16/charts-worlds-differing-view-paris-beirut-attacks (accessed 29.03.16)
2
Sunstein, C.R. (2014)
3
Hausmann, D.M. & Welch, B. (2010)

10
so or cannot be expected provide assistance. Agencies can use digital volunteers to find initiatives,
hashtags and posts that the agency can promote and repost. Crisis maps can also play a role in citizen-
to-citizen emergency management, allowing people to post offers of shelter and food and calls for
assistance, such as help to clear a road, on an interactive, online map.
By “nudging” the stream of posts on social media and facilitating citizen-to-citizen management,
emergency management agencies use social media to change the behavioural norms in emergencies to
create citizens that actively participate in emergency management. This behavioural change is not an
attempt to move the responsibility of emergency management from agencies to the citizens, but to
increase citizens’ awareness of the potential they have to assist agencies and fellow citizens in
emergencies through social media.

11
4. Crisis mapping
Emergency management agencies can use big data from social media and crowdsourcing to create
crisis maps. Crisis maps are one way to display information from social media and to encourage public
participation, and official agencies, volunteers and local organisations increasingly use crisis maps to
improve situational awareness and coordination in emergencies. In countries with many smart phone
users, people can easily geotag their posts on social media and interactive crowdmaps can locate the
user making it easy to post location specific information. Crisis maps can be simple using forms, or
complex, encompassing information from many different sources. Crisis maps can be open and
uncontrolled, or controlled and operated by emergency management agencies with the help of
volunteers.
One way of creating a crisis map is to have an online map that asks people to answer a simple question
with a yes/no answer or pre-set alternatives in an online form. For example in case of extreme weather
causing flooding, people will first confirm or plot their location on the map, then answer the question
“Is your area flooded? Yes/No” and then, if yes, be asked describe the flooding according to some
alternatives, such as “standing water in the road” or “water in the basement”. The alternatives can come
with photos in order to help people choose the right alternative(s). If enough people respond to the call
for reports, the map will give the emergency management agency a good overview over which areas
that need assistance and help the agency to prioritise the emergency response. This type of crisis
mapping is also called crowd mapping. The Swedish Civil Contingency Agency (MSB) has used crowd
mapping twice since December 2015 to test if people can hear the emergency siren. The first time MSB
tested the map it crashed because of all the activity, but the second time over 14 000 people registered
if they could hear the siren or not in an OpenStreetMap. MSB wants to use these simple tests to
develop a crisis map with several functions. Other emergency management agencies, such as the
Federal Emergency Management Agency (FEMA) in the USA and Emergency Management Australia,
have apps with interactive crisis maps where people can add photos.
Another way of creating a crisis map is to use volunteers to filter, sort and plot information from social
media and news channels in an online map. This information includes situational reports, calls for
assistance and shelter, local news and photos. The emergency management agency can also ask the
public to submit reports through texts, online forms, emails and calls. This way of making crisis maps
has been used during periods of violence, such as the violence after the election in Kenya in 2007 and
the civil war in Syria, and in the response to natural disasters, such as the Nepal Earthquake in 2015.
During periods of violence, the main purpose of the crisis maps has been to inform people about safe
and unsafe areas and to create awareness about the situation, registering reports of missing people,
violence and riots. In the response to natural disasters, the main purpose of the crisis maps has been to
map the needs of the affected population, reduce response time, and reach people in need of
assistance.
The type of emergency will affect which functions a crisis map can have, privacy requirements and
what information that can be available to the public. In periods of violence, terror or riots, the
anonymity of the people reporting incidents is important, whereas in the response to natural disasters
the ability to communicate with the people sending reports can be necessary to verify information or
get additional details. The issue of privacy versus the ability to contact the people who provide
information is a challenge that emergency management agencies have to consider. Crisis maps can be
public or private, only accessible to the agencies and actors involved in the response, or have two

12
different pages. Having a public map can encourage people to report and can also be used to request
reports on specific issues or from specific areas. In addition to online forms and SMS services
specifically connected to the crisis map, agencies can also consider using the conventional
communication channels, such as the agency’s phone number, to gather situational reports that
volunteers can sort and place in an interactive map.
Figure 3: Call for help sent to and posted on the Ushahidi map after the earthquake in Nepal in 2015.

13
How to exploit the full potential?
This memo has presented four ways using big data from social media and crowdsourcing to improve
situational awareness in emergencies. How can emergency management agencies combine several
approaches and exploit the full potential?
Automatic filters can track trends and sudden spikes in activity, and is mainly a matter of finding the
right filters for social media and creating the technical solution. The example of the U.S. Geological
Survey (USGS) shows that emergency management agencies can benefit from doing some research on
the type and nature of posts during certain emergencies. The USGS found that people tweeting about
earthquakes make short tweets and that first hand reports are unlikely to include links or numbers, so
they filter posts based on these two criteria. Emergency management agencies should consider
collaborating with universities and graduate students on finding the appropriate criteria for automatic
filters. Agencies can also consider using digital volunteers to develop filters. Automatic filters can have
great potential to map emergencies. A study by Kryvasheyeu et al. (2016) used Twitter posts with
geotagging posted during hurricane Sandy in 2012 to create an intensity map of the damage, possibly
more accurate than the current tools that the Federal Emergency Management Agency (FEMA) uses
today. Kryvasheyeu et al. (2016) combined Twitter activity with population data to match the activity
with population density. The study shows the potential for combing big data from social media with
other data to more accurately map emergencies. When creating automatic filters, agencies have to
consider how to develop the appropriate technical solution that can visualise and integrate the
information from the automatic filters into the existing operational procedures. Developing automatic
filters may require substantial resources, so agencies have to carefully consider what type of
information that can improve emergency management.
Emergency management agencies can use digital volunteers in different ways and use different types of
volunteers. Volunteers can be spontaneous, or be connected to the agency over a long period of time.
Volunteers can be untrained and placed around the world, or be trained and have a lot of knowledge
about emergency management in specific areas. In large humanitarian disasters, the large international
networks of digital volunteers can provide emergency management agencies and humanitarian
organisations with a large group of volunteers that can go through huge amounts of information and
work around the clock. However, in national emergency management, such as in Denmark, a closed
network of trusted, trained volunteers or gatekeepers may be more appropriate and more efficient than
using remote volunteers. Emergency management agencies are more likely to get the appropriate
information and support from volunteers that are familiar with the agency they are supporting and the
emergency management procedures of that agency. In addition, digital volunteers must be able to
integrate the information they retrieve with the information generated by automatic filters, and this
requires some knowledge of the technical solutions and the operation centre of the specific agency.
Emergency management agencies have to offer a lot of support to volunteers, so finding the right
solution is necessary to get “value for money” and ensure good cooperation.
Emergency management agencies should consider how they can influence the activity on social media
in emergencies. Agencies can use their official channels to “nudge” people to post more useful
information and information that is easier to track. Agencies can also use their official channels to
facilitate citizen-to-citizen emergency management. “Nudging” and facilitating citizen-to-citizen
emergency management do not necessarily require large resources, but agencies have to be present on
social media and do some research on people’s behaviour on social media in emergencies. Emergency

14
management agencies can use automatic filters and digital volunteers to make these interventions
more effective. Automatic filters can show popular hashtags, and agencies can use volunteers to
discover specific initiatives and to check the actual content of trends.
Crisis maps are one way to display information from social media and crowdsourcing. Some crisis map
platforms, such as Ushahidi, allow emergency management agencies to integrate several of the already
mentioned approaches. In these platforms agencies can collect, manage and visualise data received
through SMS, online forms, email and Twitter. The Ushahidi platform has been used in several large
emergencies, such as the Haiti earthquake in 2010. Ushahidi was then used in combination with a free
SMS service where people could send reports. Volunteers aggregated individual reports to identify
clusters of incidents and urgent needs, and gathered information from social media. All the
information was plotted on an interactive map and responders on the ground received feeds of
organised information, such as reports of injured and trapped people and messages expressing needs
for food and shelter. Crisis mapping is often presented as the way forward for emergency management,
but the value of the map depends on the quality of information put into the map and its compatibility
with existing systems. Consequently, if emergency management agencies wish to develop and use crisis
mapping, they should first consider what type of information that should feed into the map and which
type of information the map should display. Agencies do not necessarily have to invest in complex
platforms, as crowd maps asking the public to answer simple questions can give agencies a good
overview of the situation. Agencies also have to consider whether they should employ and use several
maps to display different types of information.
This memo has presented several ways of using big data from social media and crowdsourcing to
improve situational awareness in emergencies. Emergency management agencies have to consider how
they best can combine the different approaches, how they can integrate these approaches into existing
operational procedures and how they have to adapt in order to process and use these new sources of
information. Adapting to a new digital future of emergency management requires most emergency
management agencies to radically change the way they operate and, like the American Red Cross,
establish and run digital operation centres that can process, visualise and communicate the
information that is retrieved from social media and crowdsourcing. Even automatic filters, which can
easily be developed, require the necessary capacity in the emergency management agency to process
and utilise the information they provide.

15
Challenges and risks
Big data from social media and crowdsourcing have potential to improve situational awareness in
emergencies. The main challenge for emergency management agencies is to adapt to the new digital
future of emergency management and accept the increasingly important role of digital volunteers and
citizens. However, even when emergency management agencies are ready to use the information from
the public they face challenges related to the quality of the data, and the risks related to responsibility
and issues of privacy.
One of the main challenges is the amount of information that is available, the majority of which is
noise and distraction. Processing all this data is time-consuming and requires both computer and
human processing. If emergency management agencies use volunteers, the volunteers spend a lot of
time looking through useless information. Another challenge is the representativeness of the data.
How can emergency management agencies know that the data reflects reality? The absence of reports
does not necessarily mean that there is nothing going on, or the opposite, a flood of posts on Instagram
and Twitter does not necessarily mean that an area is hit most severely. Some areas might have power
blackouts that prevent or reduce activity on social media and some areas have older populations that
are not well represented online. Furthermore, people use categories and classifications differently and
do not necessarily have the same understanding of the seriousness of events as emergency
management agencies. For example, people can hashtaging their posts using “flood” when the reality is
just standing water in the road. Other challenges relating to the quality of data are misinformation,
rumours and trolling. Just hours after the attacks in Brussels fake footage of the attacks spread rapidly
across social media4
. However, false and incorrect information is also a challenge for traditional
emergency lines. Keeping all these challenges in mind, emergency management agencies must accept
the premise for using data from the public to gain situational awareness: that the overall benefit of the
system outweighs the work involved and the time spent responding to incorrect or wrong information.
Increasing the use of and interaction with people through social media and crowdsourcing create
several risks. First, when emergency management agencies become more present on social media, the
public also expects to be heard. Agencies risk that individuals use social media to ask for help when
they should use conventional emergency channels. Even if agencies use volunteers to go through the
posts on social media or promote certain hashtags for calls for help, they are likely to miss some calls
for help. Second, using social media and crowdsourcing creates risks related to privacy and the sharing
of confidential information. In the case of a man-made crisis, such as terrorism, there are risks of the
wrongful identification of suspects and the sharing of confidential police information. Both things
happened after the Boston bombings in 2013. Emergency management agencies also bear the risk of
illegally using and storing information people have posted on social media. In situations of violence,
emergency management agencies must ensure the anonymity of people submitting reports to
crowdmaps and be careful not to post or allowing people to post sensitive information in an open map.
In addition to the challenges and risks, researchers and practioneers debate whether utilising big data
from social media and crowdsourcing has any effect on emergency response.
4
http://www.theguardian.com/media/2016/mar/23/fake-youtube-videos-brussels-attacks-facebook-twitter (accessed 29.03.16)

16
A critical note
Optimistic narratives believe that big data from social media and crowdsourcing will revolutionise
emergency management. In these narratives, the challenges and risks mentioned in the previous
section are mainly seen as technical problems that can eventually be solved, or at least risks worth the
gain. However, many argue that big data from social media and crowdsourcing produce only partial
and skewed images of emergencies, and are only useful as two sources of information among many5
.
This scepticism is rooted in several concerns: the neutrality and representativeness of data from social
media, usefulness, issues of privacy and ethics, and the image big data from social media creates of
emergency management.
Social media platforms are not neutral, but shape and influence the messages that people post. For
example, people will shape their posts on Twitter and Instagram depending on their followers and
social networks. In addition, posts by some people will be reposted more often than others and quicker
reach the emergency management agencies because the account is popular and many followers.
Messages on social media may have multiple contexts and meanings, and this context is often lost
when the data is aggregated to produce trends. Furthermore, social media does not portray the full
scope of an event. The population and age groups in an area and potential damages on power supply
affect the number and locations of posts. During the hurricane Sandy in the US in 2012, most Twitter
posts came from Manhattan although other areas were more severely hit but suffered from power
outages. As the study by Kryvasheyeu et al. (2016) shows, agencies can benefit from combining data
from social media with population density data to generate more accurate maps. Emergency
management agencies always have to question whether patterns exist or if they seek patterns where
there are none.
Big data from social media and crowdsourcing can give emergency management agencies detailed
information, such as calls for help, and aggregated data in the form of trends. Many people have
praised the usefulness of both types of information. However, reports of the usefulness of this
information in the response to the Haiti earthquake were varied. Actors involved in the response in
Haiti disagreed about whether the information was useful at all, or what type of information that was
most useful6
. Case studies conducted by the EU project COSMIC7
showed that the use of social media
was very limited in many of the cases, such as the heat wave in the UK in 2013. In addition, in this case
old people were most affected by the heat wave, and also the ones less likely to be one social media.
Emergency management agencies have to consider if big data from social media is more relevant for
some emergencies than others, and what type of information that will improve situational awareness
and emergency management.
Using big data from social media and crowdsourcing to create crisis maps poses several issues of
privacy and ethics. During the response to the Haiti earthquake people used a free SMS service to
report damages, people trapped in ruins and request assistance, food and shelter. This SMS service was
eventually integrated into the Ushahidi platform used by many of the agencies participating in the
relief efforts. However, a study conducted later showed that none of the people interviewed in the
study were aware that their text messages were made public in the Ushahidi platform, but considered
5
For example, Burns (2015)
6
Hazelman & Waters (2010)
7
COSMIC will deliver a set of instructions, recommendations and best practices related to the exploitation of
social media in emergency situations. Read more at: http://www.cosmic-project.eu/

17
these text messages as private8
. Since the Ushahidi map was in English, Kreyol speakers did not have
access to the map. This study shows that emergency management agencies must be clear on how they
will use the information provided to them. After the Boston bombings in 2013, the Boston police used
social media to ask for photo or video material that could help them to identify the suspects. By doing
this they unintentionally invited the public to speculate and several people were wrongly identified as
suspects causing long-term damage for these people and their families. Emergency management
agencies must carefully consider the potential consequences of engaging with the public, and always
weight the costs and benefits of using social media in and after emergencies.
As social media platforms are public, agencies can easily assume that they can freely use the data
gathered from these platforms. However, many people do not know how to manage their own privacy
settings and are not well-informed about the extent that external agents can collect data from social
media. People’s privacy preferences will also depend on their circumstances. For people in emergency
situations the need to get assistance or to reach out to friends and loved ones can be greater than
protecting their privacy and location. By using data from several platforms, emergency management
agencies can gather personal information without people’s knowledge or consent. In order to create
useful automatic filters that can be used in emergencies, emergency management agencies have to
know what the “normal” is, i.e. create some baselines. This requires constant monitoring of social
media activity. For example, to discover if more people posts on Instagram in a village in Jutland,
Denmark, during a storm, the agency must know the number of posts on “normal” days. Hence, if
emergency management agencies want to use social media to improve situational awareness in
emergencies, they have to monitor and store a lot of data. This raises some ethical questions about
surveillance of citizens and can be illegal according to laws concerning privacy and personal
information.
Crisis mapping is one way of gathering and visualising information from social media and
crowdsourcing in emergencies, and it is increasingly popular. However, some critical scholars argue
that crisis mapping has several important implications for how we understand emergencies9
. First,
crisis mapping starts when the emergency occurs and does not include the problems and factors that
contributed to the emergency, such as poor infrastructure. Hence, crisis mapping gives a very one-side
view of the emergency and does not put the emergency in a long-term context. Second, crisis mapping
using remote digital volunteers implies that people situated far away from the emergency with no prior
knowledge about the area/country or the emergency management agency can understand the data
they collect and process. Emergency management agencies should consider this point of critique in any
situation when they use remote digital volunteers. Third, crisis mapping assumes that information
about the emergency can be categorised and aggregated. Lastly, many types of crisis mapping creates
the impression that digital volunteers are necessary for effective emergency management.
This critical note on using big data from social media and crowdsourcing in emergencies does not
imply that information from the public is useless for emergency management agencies. However, using
this information poses some serious concerns that are not merely technical problems. Emergency
management agencies must be careful when using these new sources of information and be aware of
their implications for understanding and managing emergencies.
8
Hazelman & Waters (2010)
9
For example, Burns (2015)

18
How can DEMA use big data from social media and
crowdsourcing to improve situational awareness in
emergencies?
Big data from social media and crowdsourcing have potential to improve situational awareness in
emergencies. A possible solution for DEMA is to develop a service that can be activated and offered to
other agencies in emergencies. A digital support team or a digital operations centre can be a type of
assistance that DEMA can offer to other agencies, like DEMA offers other types of assistance.
In order to move forward in developing a concept and a solution, DEMA has to consider several things:
1. Approach: DEMA has to consider which of the four approaches that fit the needs of DEMA and
other agencies. Which functions will a digital support team or digital operations centre have?
What type of information can they provide? Operational staff, IT specialists and staff from
relevant agencies should be consulted before proceeding.
2. Privacy laws and regulations: DEMA must consider the laws and regulations that can restrict
using and storing data from social media and crowdsourcing. DEMA also has to consider how
to protect people’s privacy, particularly in made-made emergencies.
3. Type and level of emergency: Should the platform apply to both man-made and natural
emergencies? Should the level of the emergency (seriousness and scale) affect if or which
functions that are activated?
4. Digital volunteers: Should DEMA use established networks of digital volunteers or create its
own network of digital volunteers? How and when should DEMA recruit new volunteers and
how should they be trained? Should DEMA use volunteers between emergencies?
5. Public/private: Should the information that DEMA gathers in an emergency be available to
the public? Should the “Mobilvarsling” app encourage people to post information?
6. Citizen-to-citizen emergency management: Should DEMA attempt to influence the activity
on social media? Should DEMA facilitate citizen-to-citizen emergency management? Should
DEMA develop guidelines for do’s and don’ts on social media in emergency?

19
Sources
American Red Cross (2012). “The American Red Cross and Dell Launch First-Of-Its-Kind Social Media
Digital Operations Center for Humanitarian Relief”. Available at: http://www.redcross.org/news/press-
release/The-American-Red-Cross-and-Dell-Launch-First-Of-Its-Kind-Social-Media-Digital-Operations-
Center-for-Humanitarian-Relief (accessed 15.02.16)
Barnes, T.J. (2013). Big data, little history. Dialogues in Human Geography, 3 (3): 297-302
Burns, R. (2015). Rethinking big data in digital humanitarianism: practices, epistemologies, and social
relations. GeoJournal, 80 (4): 477-490
Byington, G. (2015). ”American Red Cross Digital Operations Center”. Dell Blog. Available at:
http://en.community.dell.com/dell-blogs/direct2dell/b/direct2dell/archive/2015/03/31/a-year-later-
future-looks-even-brighter-for-the-american-red-cross-digital-operations-center (accessed 15.02.16)
Cobb et al. (2014). Designing for the Deluge: Understanding & Supporting the Distributed,
Collaborative Work of Crisis Volunteers. CSCW'14, February 15 - 19 2014, Baltimore, MD, USA
Crawford, K. & Finn, M. (2015). The limits of crisis data: analytical and ethical challenges of using social
and mobile data to understand disasters. GeoJournal, 18: 491-502.
Ellis, E. (2015). ”How the USGS uses Twitter data to track earthquakes”. #DataStories, Twitter blog.
Available at: https://blog.twitter.com/2015/usgs-twitter-data-earthquake-detection (accessed 07.03.16)
Hausmann, D.M. & Welch, B. (2010). To nudge or not to nudge. Journal of Political Philosophy, 18 (1):
123-136
Heinzelman, J. & Waters, C. (2010). Crowdsourcing Crisis Information in Disaster Affected Haiti.
Special Report 252, the United States Institute of Peace.
iRevolutions (2014). “The Filipino Government’s Official Strategy on Crisis Hashtags”. Available at:
http://irevolutions.org/2014/07/01/filipino-official-strategy-crisis-hashtags/ (accessed 29.03.16)
iRevolutions (2013). ”Social Media for Emergency Management: Question of Supply and Demand”.
Available at: http://irevolutions.org/2013/04/23/smem-supply-demand/ (accessed 29.03.16)
Jackson, J. (2016). ”Fake Brussels YouTube videos prove ease of digital communication”. The Guardian.
Available at: http://www.theguardian.com/media/2016/mar/23/fake-youtube-videos-brussels-attacks-
facebook-twitter (accessed 29.03.16)
Krisinfobloggen (2013). ”Crowdsorucing: när många loser problem tilsammans”. Available at:
https://blogg.msb.se/krisinformation/?s=crowdsourc (accessed: 07.03.16)
Kryvasheyeu et al. (2016). Rapid assessment of disaster damage using social media activity. Science
Advances, 2 (3)
Merrick, D.F. & Duffy, T. (2013). Utilizing Community Volunteered Information to Enhance Disaster
Situational Awareness. ISCRAM 2013 Conference, May 12-15, Baden-Baden, Germany

20
Munkvold et al. (2015). Volunteers’ Perceptions of Social Media in Emergency Management. ISCRAM
2015 Conference, May 24-27,Kristiansand, Norway.
Nuwer, R. (2016). “Twitter May Be Faster Than FEMA Models for Tracking Disaster Damage”.
Smithsonian.com. Available at: http://www.smithsonianmag.com/science-nature/twitter-may-be-
faster-fema-models-tracking-disaster-damage-180958391/?no-ist (accessed 29.03.16)
Noula.ht crisis map. Available at: http://www.noula.ht/ (accessed 23.02.16)
Papadimitriou et al. (2013). Case studies of communication media and their use in crisis situations.
COSMIC deliverable D2.2.
Ser, K.K.S. (2015). ”In charts, the world’s differing view of the Paris, Beirut attacks”. PRI. Available at:
http://www.pri.org/stories/2015-11-16/charts-worlds-differing-view-paris-beirut-attacks (accessed
29.03.16)
Simon et al. (2015). Socialising in emergencies – A review of the use of social media in emergency
situations. International Journal of Information Management, 35: 609-619.
Sunstein, C.R. (2014). Nudging: A Very Short Guide. Journal of Consumer Policy, 37 (4): 583-588
Ushahidi. “ Powerful features”. Available at: https://www.ushahidi.com/features (accessed 15.02.16)
Ushahidi. “Quakemap”. Available at: https://www.ushahidi.com/case-studies/quakemap (accessed
15.02.16)
Verity, A. (2011). OCHA’s Lessons Learned. Collaborating with V&TCs for Libya and Japan.
Communities of Interest, 15-June Wash-up, New Work, USA.
Yin et al. (2012). Using Social Media to Enhance Emergency Situation Awareness. IEEE Intelligent
Systems, 27(6), 52-59

21
Figures
Figure 1: https://blog.twitter.com/2015/usgs-twitter-data-earthquake-detection (accessed 07.03.16)
Figure 2: http://www.redcross.org/news/article/tx/dallas-fort-worth/Red-Cross-and-Dell-Open-North-Texas-
Social-Media-Listening-Center (accessed 07.03.16)
Figure 3: Source: https://www.ushahidi.com/case-studies/quakemap (accessed 15.02.16)

Big Data from Social Media and Crowdsourcing in Emergencies

More Related Content

What's hot (19)

Viewers also liked (20)

Similar to Big Data from Social Media and Crowdsourcing in Emergencies (20)

Recently uploaded (20)

Big Data from Social Media and Crowdsourcing in Emergencies