Authors: Andrea H. Tapia Kathleen A. Moore  Nicholas J. Johnson Penn State University College of Information Sciences and Technology

1. Beyond the Trustworthy Tweet:
A Deeper Understanding of Microblogged Data Use
by Disaster Response and Humanitarian Relief
Organizations
Andrea H. Tapia
Kathleen A. Moore
Nicholas J. Johnson
Penn State University
College of Information Sciences and Technology

2. Opportunity:
Microblogging
New technological affordances allow for…
• data from crowds of non-professional participants
• average citizens reporting on activities on-the-ground during
a disaster
• a way of mitigating the impacts of and speeding up the
recovery from extreme events
• is ubiquitous, rapid and accessible
• empowers average citizens to become more situationally
aware during disasters
• coordinate to help themselves

3. Problem Statement: Despite the evidence of strong value to
those experiencing the disaster and those seeking information
concerning the disaster, there has been very little uptake of
message data by large-scale, disaster response organizations
Research Questions:
(1) What are the criteria for measurement, standards and threshold for
relevant, trustworthy and actionable data to disaster response
organizations?
(2) How can automatic measures of trust be incorporated into
organizational decision-making practices so that the serving of these
data occurs at the appropriate time, in the appropriate form, to the
appropriate person, and the appropriate level of confidence?

4. Research Goal
• Map the information needs and
flow through response
organizations
• Map patterns of decisions
made during a disaster
response, types and forms of
data inputs to those decisions,
and varying standards for
relevance and veracity for each
• Understand the data
requirements at decision points
during a response

5. Old History
ISCRAM 2011—
• We reported that microblogged data produced by citizens were akin to
food that responding organizations could not eat
• Responding organizations saw the data as untrustworthy, they could not
be inserted into the critical decision tree of the organization
Data quality was the
single most important
determining factor in use

6. We considered these to be three discreet options, with larger community
favoring the third, most technical choice
However, after our second round of data collection, we see these three
potential paths all playing out in some fashion and often overlapping
Our second round of data collection has shown us that informational
needs of humanitarian organizations responding to a crisis are
varied, and standards for quality of that data also varies
ISCRAM 2011
We offered three potential paths
toward increased microblogged
data use by humanitarian
organizations:
1. bounded microblogging environment
2. ambient or contextual use
3. computational solutions to
automating trustworthiness

7. New Story
• Beyond Data Quality
• Beyond Trustworthiness
• Fast, Good Enough Data
• From a community
The landscape of the use of microblogged data in crisis
response is varied, with pockets of use and acceptance
among organizations
Microblogged data is useful to responders in situations where
information is limited, such as at beginning of an emergency
response effort, and when risks of ignoring an accurate
response outweigh risks of acting on an incorrect one

8. THE GOOD ENOUGH PRINCIPLE IN
HUMANITARIAN ACTION
• There have always been imperfect data and knowledge
during disasters
• Act on good intelligence, not perfect
• Satisficing and good enough principles
• Strive for helpfulness, not accuracy

9. RESEARCH DESIGN
Our Team: EMERSE: Enhanced Messaging for Emergency
Response
Facilitated by our partner NetHope.org. Second Round of Data
Collection.
• Interviewed representatives from 21 of the 38
NetHope members
• Information managers associated with Emergency
Response
• Applied analytical induction coding process

11. Findings: A Varied Landscape of Data Quality
Summary: Responders already make decisions based on imperfect data, often
from second-hand sources. The inherently chaotic nature of any disaster limits
responders’ ability to both gather and assess the quality of information from
traditional sources.
Subjects said:
• We use best data available, but in most cases high quality data is never
available
• We regret that better data was not typically available, acceptance that this
condition was part of the nature of their work, and understanding that despite
lack of perfect data hundreds of emergencies had been responded to
successfully, millions of lives had been saved and regions had been
reconstructed

12. Findings: A Varied Landscape of
Decisions
Summary: Information needs changed as
the disaster environment changed
Onset of a disaster needed to
understand context and scope of an
emergency, including size and location of affected
population and extent of damage to basic support
infrastructure
Later, they need information about specific gaps in availability of goods, services and
other forms of aid
Still later, they need information about operational coordination, i.e. who is
responding with what and where
Lastly, they need regular updates on the security situation, impact of intervention,
status of affected population, and constant inter-organizational coordination of
information
Microblogged data’s value as an information source is not a constant, and
would vary as a disaster response develops.

13. Findings: Data Type Influences the
Required Level of Data Quality
Summary: Requirements for data quality and trustworthiness
were variable depending on the type of question asked by the
responding organization
• initial awareness of a disaster they would accept a low or unknown
threshold for data quality in exchange for real time knowledge
• looked to social media data during first few days after a disaster for
contextual data
• Around half of subjects stated that they would listen to microblogged
data if it spoke of a security threat to NGO field workers, supplies or
camps
• some types of questions that required a very high level of confidence
in which microblogged data could not yet be used as a key input

14. Findings: Networks of Responders Cross Organizational
Boundaries
Summary: All subjects followed members of the humanitarian community
via social media.
• Each had a patchwork of different sources of microblogged data including
official accounts, unofficial and informal blogs of employees of these
organizations, employees of various organizations, blogs of humanitarian
focused or interested individuals and family and friends
• Personal social network was already producing trustworthy and
actionable data
• It served as a powerful informal source of information about the response
and the conditions during a disaster

15. Findings: Reliance on Volunteer and Technical
Communities
Summary: Subjects mentioned that they had already used
or were planning to use secondary microblogged data--
Twitter data that had been collected on a large scale and
processed by outside groups
• Expressed more trust of volunteer and technical communities
than original data-- Ushahidi, Crisis Mappers, The Standby
Task Force and the Digital Humanitarian Network
• They transferred expertise and trust to outside their
organization, to trusting volunteers processing data rather than
data itself

16. Discussion: Organic Bounded Trust Community
• A form of bounded environment has in part grown organically
• Employees and volunteers already working in relief sector have
become active social media users, perhaps overcoming a
technology adoption problem as previously suggested
• Participants in these networks are friends and friends of friends and
largely trusted, sharing same cultural understandings of
humanitarian response and practice
• This is largely informal, organic and crosses organizational borders
and hierarchies
• The data produced in these groups serves as supplemental input to
decisions made by organizational responders

17. Conclusions:
• Twitter is a food that has always
been consumed by response
community, but in varied forms and
times, which may or may not be
official or formal channels
• Encourage response organizations
to recognize their own current use of
microblogged or crowdsourced data
and validate that use with additional
organizational suppor
• Adjust organizational standards for
data quality and accuracy based on
type of decision

18. Encourage humanitarian
workers to use social media
regularly and build networks
organic network of humanitarian
microblogging users could serve as
a middle ground between traditional
data sources and unfiltered
microblogging data
• Non-competitive nature of goals of humanitarian response organizations
is ideal for fostering an environment for inter-organizational information
sharing
• Having an informal, everyday knowledge of what other humanitarian
workers are doing could lead to better organizational efficiency and co-
ordination of response and recovery efforts
• With use comes normalcy and trust

19. • Encourage response organizations to continue to pursue
computational and automatic solutions assessing accuracy
and trust in crowdsourced data
• Shift some of burden to outside volunteer organizations
• Ensuring that organizations that specialize in processing
microblogged data are producing reliable data could also
help encourage humanitarian response organizations to
use these sources

20. Thank you.
Questions?