Conference PaperPDF Available

A Framework for Strengthening Data Ecosystems to Serve Humanitarian Purposes

Authors:

Abstract and Figures

The incidence of natural disasters worldwide is increasing. As a result, a growing number of people is in need of humanitarian support, for which limited resources are available. This requires an effective and efficient prioritization of the most vulnerable people in the preparedness phase, and the most affected people in the response phase of humanitarian action. Data-driven models have the potential to support this prioritization process. However, the applications of these models in a country requires a certain level of data preparedness. To achieve this level of data preparedness on a large scale we need to know how to facilitate, stimulate and coordinate data-sharing between humanitarian actors. We use a data ecosystem perspective to develop success criteria for establishing a "humanitarian data ecosystem". We first present the development of a general framework with data ecosystem governance success criteria based on a systematic literature review. Subsequently, the applicability of this framework in the humanitarian sector is assessed through a case study on the "Community Risk Assessment and Prioritization toolbox" developed by the Netherlands Red Cross. The empirical evidence led to the adaption the framework to the specific criteria that need to be addressed when aiming to establish a successful humanitarian data ecosystem.
Content may be subject to copyright.
A Framework for Strengthening Data Ecosystems to Serve
Humanitarian Purposes
Elise Haak
Delft University of Technology
The Netherlands
haakelise@gmail.com
Jolien Ubacht
Delft University of Technology
The Netherlands
j.ubacht@tudelft.nl
Marc Van den Homberg
The Netherlands Red Cross
The Netherlands
marcjchr@gmail.com
Scott Cunningham
Delft University of Technology
The Netherlands
s.cunningham@tudelft.nl
Bartel Van den Walle
Delft University of Technology
The Netherlands
b.a.vandewalle-1@tudelft.nl
ABSTRACT
The incidence of natural disasters worldwide is increasing. As a
result, a growing number of people is in need of humanitarian
support, for which limited resources are available. This requires an
eective and ecient prioritization of the most vulnerable people
in the preparedness phase, and the most aected people in the
response phase of humanitarian action. Data-driven models have
the potential to support this prioritization process. However, the
applications of these models in a country requires a certain level of
data preparedness. To achieve this level of data preparedness on a
large scale we need to know how to facilitate, stimulate and coor-
dinate data-sharing between humanitarian actors. We use a data
ecosystem perspective to develop success criteria for establishing a
“humanitarian data ecosystem”. We rst present the development
of a general framework with data ecosystem governance success
criteria based on a systematic literature review. Subsequently, the
applicability of this framework in the humanitarian sector is as-
sessed through a case study on the “Community Risk Assessment
and Prioritization toolbox” developed by the Netherlands Red Cross.
The empirical evidence led to the adaption the framework to the
specic criteria that need to be addressed when aiming to establish
a successful humanitarian data ecosystem.
CSS CONCEPTS
Information systems Information systems applications;
data warehouses; data analytics;
Applied computing Opera-
tions research;
forecasting;
Social and professional topics
Professional topics;
management of computing and information
systems;
KEYWORDS
humanitarian sector, data ecosystem, data preparedness, gover-
nance, framework
Permission to make digital or hard copies of part or all of this work for personal or
classroom use is granted without fee provided that copies are not made or distributed
for prot or commercial advantage and that copies bear this notice and the full citation
on the rst page. Copyrights for third-party components of this work must be honored.
For all other uses, contact the owner/author(s).
dg.o ’18, May 30-June 1, 2018, Delft, Netherlands
©2018 Copyright held by the owner/author(s).
ACM ISBN 978-1-4503-6526-0/18/05.
https://doi.org/10.1145/3209281.3209414
ACM Reference Format
E. Haak, J. Ubacht, M. van den Homberg, B. van der Walle, S. Cun-
ningham. 2018. A Framework for Strengthening Data Ecosystems to
Serve Humanitarian Purposes. In Proceedings of 19th Annual In-
ternational Conference on Digital Government Research (dg.o’18),
Anneke Zuiderwijk and Charles C. Hinnant (Eds.). ACM, New York,
NY, USA, 9 pages.
1 DATA FOR THE HUMANITARIAN SECTOR
The humanitarian sector worldwide is under pressure: there is an
increasing demand for humanitarian aid and emergency response
services, as both the number and length of humanitarian crises
grow [
1
,
19
,
33
]. One of the causes of this increase is the rising
number of disasters triggered by natural hazards, due to climate
change, increased urbanization and population growth. This results
in a fast-growing number of people in need [
12
]. Due to scarcity of
budget and resources for disaster response, humanitarian actors can-
not reach all aected persons. The prioritization of vulnerable and
aected people is therefore at the core of all humanitarian interven-
tions, and equally important in both the preparedness (pre-disaster)
and the response phase (post-disaster) of humanitarian action [
30
].
Obtaining reliable and objective information for decision-making
related to this prioritization process is considered very challeng-
ing [
29
]. Contributing factors are for example aid workers operating
under stress and time pressure and being presented with uncer-
tain information, which could lead to decision-making biases [
7
],
and the complex organizational humanitarian setting in which
information is scattered across many dierent organizations and
sectors [20].
The rapidly changing information environment provides ways
to deal with these challenges. The fast increase in the availability
of data leads to a shift towards more evidence-based humanitarian
decision-making, oering humanitarian actors ways to be become
more eective and ecient. Paper-based processes are increasingly
digitized and trends around crowdsourcing, social media, collabora-
tive digital spaces and mobile services lead to a signicant increase
in data availability. This also aects the humanitarian sector at
large. Humanitarian donors for example, push the organizations
they sponsor to open and share their data. In addition, major global
agreements that recently came into eect, such as the Paris Climate
Agreement, the Sustainable Development Goals and the Sendai
Framework for Disaster Risk Reduction, stimulate countries to col-
lect data on dierent indicators that are potentially valuable for
humanitarian organizations [
6
,
30
]. As Sabou [
27
, par. 3] describes:
“[H]umanitarian organizations now seek better ways to collect, store
and use digital data to improve their collective responses to large scale
crises” The United Nations term this trend “the data revolution” [
13
].
As a result, a growing number of studies is dedicated to mathe-
matical models and data-analytic techniques that support the hu-
manitarian decision-making processes in the preparation for or
response to the consequences of a disaster [
21
]. Data-driven pri-
ority indices have grown popular for eectively and eciently
identifying communities most vulnerable for or aected by dis-
asters and numerous data and machine learning tools have been
developed to aid prioritization [5].
However, humanitarian actors miss timely, reliable and sucient
granular data and often lack the skills and tools to analyze the data
needed for a transparent and structured prioritization [
30
]. Data
is collected, stored and used by an uncoordinated, diverse group
of actors with varying degrees of expertise. This lack of alignment
causes a range of ineciencies [
6
]. In addition, the collection, colla-
tion and analysis of data often only starts once a disaster hits, and
to be able to apply these prioritization models, a certain level of
‘data preparedness’ is required [
30
]. Data preparedness is dened
by Raymond & Al Achkar [25, p. 3] as “the ability of organizations
to be ready to responsibly and eectively deploy data tools before a
disaster strikes”. Van den Homberg, Visser & Van der Veen [
30
] go
beyond only the aspect of being ready to deploy tools and include
the pre-staging of data. They dene Data Preparedness as all ac-
tivities, that can be done before a disaster hits, to pre-stage data
with suciently high data quality (that matches the prospective
information needs of responders- like data on population density
and vulnerability, (types of) roads, house construction material,
sanitation and existing hospitals) and to develop capacities to col-
lect data on aected communities and areas once a disaster hits
to ensure a timely, ecient, and eective response. Their frame-
work for Data Preparedness consists of ve components, being
’Data Sets’, ’Data Services and Tooling’, ’Data Literacy’, ’Data Gov-
ernance’ and ’Networked Organizations for Data’. Consequently,
there is a need to understand how to facilitate, stimulate and coor-
dinate data-sharing between humanitarian actors to increase the
level of data preparedness in a country and hence be able to apply
data-driven models that can help to come to a better-informed pri-
oritization of humanitarian aid. We address this knowledge gap by
taking a “data ecosystem” approach, thereby referring to “the peo-
ple and technologies collecting, handling and using the data and the
interactions between them” [
22
, p. 557]. By applying this approach
to the humanitarian sector, we developed a systematic approach for
humanitarian organizations to develop humanitarian data ecosys-
tems that optimize the availability of data for prioritization and
decision making.
The ‘Data Services’ component of the above mentioned frame-
work, relating to the required IT infrastructures to facilitate data
preparedness, and the possible architectures of these infrastruc-
tures, will not be considered in this study as we focus on the upper
layers of the OSI model for information systems. We do recognize
that excluding the issue of the availability and design of the underly-
ing technical information architecture for storing, communicating
and analyzing the data itself needs to be considered when actu-
ally setting up a data ecosystem. The information architecture will
depend on local circumstances and readily available information
systems such as spatial data platforms, open street map, etc.
This paper is structured as follows. First, we elaborate on the
data ecosystem approach in section 2. We then present our research
design in section 3. The outcomes of a systematic literature review
to develop the generic framework for establishing a data ecosystem
is presented in section 4. In section 5, we present the evaluation
of the framework by means of expert interviews. In section 6 the
framework is adapted to t with the challenges of developing data
ecosystems within the humanitarian sector. We conclude with a
discussion of the framework and oer future research to test the
framework in a diversity of countries to further enhance its practical
use.
2 DATA ECOSYSTEM APPROACH
Academics in information intensive, socio-technical contexts have
applied the ecosystems perspective to get an idea of the diverse
interrelationships between data users, data providers, data itself,
institutions and material infrastructure [
10
], as the perspective can
help to deal with complexity [
11
]. According to Harrison, Pardo
& Cook [
10
, p. 905], the approach can be used to outline existing
conditions and develop desired conditions, stating that “its users
often aim to provoke new thinking about the conditions and require-
ments necessary to actively cultivate development of an ecosystem to
achieve a set of specic and desirable goals”. Data ecosystems can
be used “as a means for decision-making and planning” [
35
, p.18],
to locate the “relative positions of the actors in the ecosystem (data
providers, sources, resources and users)” [
31
, p. 72], and to “facilitate
access to sharing and using data” [32, p. 286].
Berens et al. already propose the use of a data ecosystem perspec-
tive on the use of digital data in the humanitarian sector: “[T]hat
of a complex data ecosystem comprised of a variety of actors that
are touched by ows of digital data due to data sharing and data-
related service delivery, and hence become part of a phenomenon
that stretches beyond their organizational boundaries” [
6
, p. 5]. The
perspective allows for a cross-organizational understanding of data
use, rather than only assessing internal data life cycles [
6
,
26
]. Ray-
mond et al. [
26
] describe the ‘humanitarian data ecosystem’ as the
network of humanitarian organizations, aliated and aected com-
munities who are producing, collecting and analyzing digital data.
We argue that it is essential that a humanitarian data ecosystem
also comprises actors other than humanitarian organizations: also
governmental and private sector organizations must be included,
given that they hold key data. In section 6 we present the roles and
responsibiliteis of stakeholders in a humanitarian data ecosystem
in detail.
With the rise in data availability from a multitude of sources
and a growing sophistication in data modeling, the data ecosystem
approach becomes increasingly important to optimize the actual
use of the data to address challenges in the humanitarian sector.
However, only limited use is made of it as of yet. Based on a litera-
ture search in the literature databases Scopus, Web of Science and
Google Scholar for ‘humanitarian data ecosystem’ or ‘humanitar-
ian’ and ‘data ecosystem’ did not yield any results. No academic
Figure 1: General framework of criteria for a successful data ecosystem
literature so far (May 2017) addresses the humanitarian data ecosys-
tem, or the success criteria for the establishment of a humanitarian
data ecosystem. The terms can be found in a limited number of
non-academic publications, but in none of these the development
of a successful humanitarian data ecosystem is addressed. Berens
et al. [
6
] and Raymond et al. [
26
] for example, discuss the humani-
tarian data ecosystem from the angle of the responsible use of data
and argue that the data ecosystem perspective can help to address
the issue of a lack of centralized governance that results in a risk of
harmful data use in the humanitarian sector. Sabou shares this view,
claiming that “the existing ‘humanitarian—digital data’—ecosystem
is essentially a collection of ungoverned pilot programs, and a new
lens is needed to understand how we can meet the increasing demand
for humanitarian aid in disaster onsets with relevant and ethical hu-
manitarian innovation.[
27
, par. 7]. Our literature review presents
an academic knowledge gap to which we contribute by developing
a framework for humanitarian organizations to develop such data
ecosystems. In the next section, we present our research design
towards this objective.
3 RESEARCH DESIGN
Our research design consisted of three phases. In the rst phase, we
used a literature review to retrieve categories of aspects that need to
be considered to set up a successful data ecosystem in general. With
these categories, we developed a framework for the success criteria
for the development of a data ecosystem. This generic framework
was evaluated with 17 academic researchers. In the second phase,
we tested this generic framework by means of a case study in order
to adapt it to the specic application in the humanitarian sector. The
case study is the Community Risk Assessment and prioritization
toolbox, developed by the Netherlands Red Cross. In the third phase,
we assessed the practical relevance of the framework by means of
interviews with eight humanitarian data experts. In the following
sections, we go into more details on these research phases, to show
the intermediate ndings towards our nal framework.
4 PHASE 1 - GENERIC DESIGN FRAMEWORK
FOR DATA ECOSYSTEMS
To start with, we performed a systematic literature review, in which
the outcomes from publications addressing the establishment of
data ecosystems in general were combined. This allowed us to
retrieve a rst set of important aspects to consider when aiming to
set up a successful data ecosystem.
Relevant literature was found by means of a thorough search
for journal and conference articles, books, reports and other infor-
mative documents. Scopus, Web of Science, Google Scholar and
Google Search were consulted as online databases. We used the
following (combinations of) keywords: ‘Data ecosystem’; ‘Open
data ecosystem’; ‘Government ecosystem’; ‘Criteria’ + ‘data ecosys-
tem’; ‘Element” + ‘data ecosystem’; ‘Successful’ + ‘data ecosystem’;
‘Establish’ + ‘data ecosystem’ ; ‘Data infrastructure’.
The search of relevant articles was an iterative process. By in-
serting (combinations of) these keywords, a multitude of articles
was found. The found articles were judged on relevance by reading
the titles and summaries, and by scanning the text: did they indeed
address criteria for developing a successful data ecosystem? When
an initial selection was made, an additional search for literature
took place based on the references used in these selected articles,
which resulted in an increase of articles with potential. The new
articles were scanned in the same manner, and by doing so we
narrowed down the set of articles to 17 publications, which are
listed in Table 1.
Next, we manually coded all criteria that were presented in
the selected articles, which led to the creation of the overview as
presented in table 2. The literature review yielded a diverse set of
aspects that needs to be considered when establishing a successful
data ecosystem. We found that all aspects could be grouped into one
of the categories identied by Welle Donker & Van Loenen [
32
], in
their study on how to assess the success of the open data ecosystem.
We therefore used this same categorization to organize the data
ecosystem success criteria resulting from our literature review.
Three dierent categories are distinguished:
Author(s) Year Focus Objective Domain of Author(s) Ref.
Attard, Orlandi & Auer 2016
Economic data ecosystem
To project their vision of generating a new Economic
Data Ecosystem that has the Web of Data as its core.
Enterprise Information Systems [3]
Barthélemy 2016 Open data ecosystem
To provide an overview of the Belgian open data ecosys-
tem
School of Management: MSc in Business
Engineering
[4]
Davies 2012
Open data infrastructures
and ecosystem
To highlight some of the interventions that may be nec-
essary to support realization of impact from open data
initiatives.
Open Data & Open Government [8]
Dawes, Vidiasova &
Parkhimovich
2016
Open government data
ecosystem
To develop a preliminary ecosystem model for planning
and designing Open Government Data programs
E-governance / Open Government [9]
Harrison, Pardo & Cook 2012
Open government ecosys-
tem
To create a research and development agenda with ques-
tions essential to the development of Open Government
Ecosystems.
IT in Government / Open Government [10]
Heimstädt, Saunderson &
Heath
2014 Open data ecosystem
To identify a set of structural business ecosystem prop-
erties.
Open (Government) Data [11]
Immonen, Palviainen &
Ovaska
2014
Open data based business
ecosystem
To dene the requirements of an open data based busi-
ness ecosystem (an open data ecosystem from the busi-
ness viewpoint)
Service / Software Engineering [14]
Jetzek 2017 Open data ecosystem
To explore the possibilities for sustainable value genera-
tion in the Open Data Ecosystem
IT Management: Big Data & Open Data [16]
Lee 2014 Open data ecosystem
To specify a series of specic elements critical for build-
ing an Open Data Ecosystem
Linked & Open Data [17]
Macharis & Crompvoets 2014
Spatial data infrastruc-
ture
To evaluate development scenarios for the spatial data
infrastructure for Flanders
Supply Chain Management / Public Gov-
ernance / Spatial Data Infrastructures
[18]
Parsons et al. 2011 Science data ecosystem
To success several short- and long-term strategies to fa-
cilitate a socio-technical evolution in the overall science
data ecosystem
Science Data [22]
Pollock 2011 Open data ecosystem
To stress the importance of data cycles with feedback
loops when building open data ecosystems
Open Knowledge [23]
Ponte 2015 Open data ecosystem
To provide an overview of the issues to be addressed
when enabling an open data ecosystem
Organization Science & ICT [24]
Van Schalkwyk,
Willmers & McNaughton
2016 Open data ecosystem
To consider the supply, demand and use of open data, as
well as the roles of intermediaries, using an ecosystem
approach
Open data / Open ICT Ecosystems &
Scholarly Communication
[31]
Welle Donker & Van Loe-
nen
2017 Open data ecosystem
To develop an open data assessment framework based
on three output indicators as conditions for a successful
open data ecosystem
Open Data / Information Infrastructures
[32]
Wiener et al. 2016 Open data ecosystem
To enable an open data ecosystem for the neurosciences
Neuroscience & Data-sharing [34]
Zuiderwijk, Janssen &
Davis
2014 Open data ecosystem
To provide an overview of essential elements of open
data ecosystems for enabling easy publication and use
of open data
Open Data / ICT & Governance [35]
Table 1: List of articles reviewed (objectives and methodologies have been retrieved from the articles)
(1) Data supply, relating to the provision of data as open data,
(2)
Governance, being the framework of policies, processes and
instruments to realize common goals in the interaction be-
tween entities (and facilitating the data supply), and
(3) User characteristics [32].
An overview of the articles in which each criterion appeared, can
be found in Table 2. Later in this section, all criteria are described
in full.
We evaluated the ndings from our literature review in a vali-
dation session with a group of open data researchers. During this
session, the completeness of the literature list (as presented in Ta-
ble 1), was assessed, as well as the completeness and correctness
of the set of data ecosystem governance criteria. All participants
were provided with a hand-out of both lists. The outcomes were
elaborated on in a presentation, after which an open discussion
took place in which the participants provided feedback. Taking this
feedback into account, the ndings from the literature review and
the validation session led to the framework as visualized in Figure 1.
In the following paragraphs we explain the criteria that are used in
the framework.
4.1 Data Supply (DS) Criteria
The Data Supply (DS”) criteria are described as follows:
DS1 Data availability and accessibility:
It should be clear for
actors in a data ecosystem how to nd data, where to nd it and
how to access it.
DS1a Data collection:
It should be clear to actors in the data
ecosystem how data can be collected.
DS1b Privacy:
It should be ensured that data protection laws
are followed.
DS1c Licensing and legal conditions:
Another important as-
pect to enhance the accessibility of data in a data ecosystem
is the type of licensing associated with the data, which is
necessary to ensure the legal foundation for the potential
(re)use of data
DS1d Aordability:
The data shared in the data ecosystem
should be aordable.
DS1e Metadata:
The existence of appropriate metadata can
help to improve the availability and accessibility of data.
DS2 Data usability:
High quality of the data shared in a data
ecosystem should be ensured to enhance its usability:
Category Criterion Description Source(s)
Data Supply
Data availability and accessi-
bility
It should be clear for actors in a data ecosystem how to nd data, where to nd it
and how to access it
[3, 9, 10, 14, 17, 18, 22, 31, 32, 34, 35]
Data collection It should be clear to actors in the data ecosystem how data can be collected Added after expert validation
Privacy It should be ensured that data protection laws are followed [9, 14, 17, 34]
Licensing and legal conditions
Another important aspect to enhance the accessibility of data in a data ecosystem
is the type of licensing associated with the data, which is necessary to ensure the
legal grounding for the potential (re)use of data
[4, 9, 14, 17, 18, 22, 31, 32, 35]
Aordability The data shared in the data ecosystem should be aordable. Added after expert validation.
Metadata
The existence of appropriate metadata can help to improve the availability and
accessibility of the data
[8–10, 17, 18, 22, 32, 34, 35]
Data usability: Data quality
High quality of the data shared in a data ecosystem should be ensured to enhance
its usability
[3, 4, 8–10, 14, 17, 18, 22, 31, 32, 35]
Data standards
An important aspect that determines the quality of data is the presence of stan-
dards to facilitate data interoperability
[4, 9, 10, 14, 17, 18, 32, 34]
Data Governance
Vision, communication and
stimulation
A collaborative, interactive environment should be established and cooperation
between stakeholders should be stimulated
[3, 4, 9, 14, 32, 34]
Division of roles and responsi-
bilities
There should be a clear division of the roles and responsibilities of the actors in
an ecosystem
[4, 8–10, 14, 17, 31, 35]
Feedback
The data ecosystem should include feedback mechanisms to enable data users to
provide feedback to data providers
[4, 9, 10, 14, 17, 35]
Leadership and incentivization
Activities in an ecosystem should be stimulated, incentivized and coordinated
(either top-down or bottom-up, depending on the context)
[4, 8–10, 14, 17, 32, 34]
Participatory capacity
Data supply should match data demand in an ecosystem; public bodies require
certain capacities to be able to participate in an ecosystem
[3, 9, 17, 32, 34]
Sustainability: nancing and
value creation
For a data ecosystem to become sustainable, sustainable nancing should be
arranged and the value should be generated for the ecosystem stakeholders
[8, 10, 11, 16, 17, 22, 24, 31, 32, 34]
Trust and transparency
Data providers and data users should mutually trust and be transparent to each
other.
Added after expert validation.
User Characteristics User capabilities The capabilities of the data users in an ecosystem should be considered [4, 8–10, 17, 32, 35]
User needs The needs of the data users in an ecosystem should be considered [9, 10, 14, 17, 32]
External Context - The external context of the data ecosystem should be considered [4, 9, 10, 18, 22, 31, 35]
Table 2: Criteria for a successful data ecosystem found in the systematic literature review, complemented with insights from expert.
DS2a Data standards:
An important aspect that determines
the quality of data is the presence of standards to facilitate
data interoperability.
DS2b Data quality:
The data shared in the data ecosystem
should be of high quality
4.2 Governance (G) Criteria
The Governance (G”) criteria are described as follows:
G1 Vision, communication and stimulation:
A collaborative,
interactive environment should be established and cooperation
between stakeholders should be stimulated. A fragmented approach
should be avoided.
G1a Division of roles and responsibilities:
All actors in the
data ecosystem should be identied, and there should be a
clear division of the roles and responsibilities of the actors
in a data ecosystem.
G1b Feedback:
The data ecosystem should include feedback
mechanisms to enable data users to provide feedback to data
providers.
G2 Leadership and incentivization:
Activities in a data ecosys-
tem should be stimulated, incentivized and coordinated by a prob-
lem owner, either top-down or bottom-up. Incentivization can take
place by for example make data sharers feel intrinsically rewarded,
or by lowering the barrier to entry for data sharing.
G3 Participatory capacity:
Data supply should match data de-
mand in a data ecosystem: public bodies require certain capacities
to be able to participate in a data ecosystem, like:
Technical knowledge on certain systems and technologies
involved
Data management knowledge, on how to ensure high data
quality
Operational knowledge, on how to incorporate data activities
into current practices
G4 Sustainability:
For a data ecosystem to become sustainable,
sustainable nancing should be arranged and value should be gen-
erated for the ecosystem stakeholders
G4a Financing: Arrange for sustainable nancing
G4b Value creation:
Value generation for the ecosystem stake-
holders
G5 Trust and transparency:
Data providers and data users should
mutually trust and be transparent to each other
4.3 User Characteristics (UC)
UC1 User capabilities:
The capabilities of the data users in an
ecosystem should be considered.
UC2 User needs:
The needs of the data users in an ecosystem
should be considered.
4.4 External Context (EC)
One last aspect related to data ecosystems coming forward in the
literature is the importance of considering the external context
of an ecosystem. This is not exactly a criterion for a successful
data ecosystem; it is merely an overarching factor that should be
considered when analyzing a data ecosystem. The functioning of a
data ecosystem also depends on for example the local culture, the
political system and historical inuences. These aspects all together
form the institutional conditions in which the ecosystem is or needs
to be embedded, that inuence how actors in the ecosystem function
and how dierent ecosystem elements are arranged. Hence:
External context:
The external context of the data ecosystem
should be considered.
As can be seen in Table 1, 14 of the 16 articles focus on open data
ecosystems and many of the authors have a background in open
(government) data. To assess whether the criteria that hold for the
successful establishment of open data ecosystems specically can
also be translated into the humanitarian data ecosystem, we used
a case study in the humanitarian sector in the next phase of our
research.
5 PHASE 2 - PRACTICAL VALIDATION
In our literature review we discovered that the ecosystem approach
was mainly applied in the research domain of open data. The litera-
ture yielded an initial set of criteria, but also raised the issue whether
these are fully applicable to the humanitarian sector, in which much
of the data is not shared openly (and sometimes not even stored
digitally). Therefore, we empirically validated the generic frame-
work of success criteria for the development of a data ecosystem in
the humanitarian sector.
For this empirical validation, we tested the framework by means
of the ‘Community Risk Assessment and Prioritization toolbox’,
developed by the data team of the Netherlands Red Cross. This is
an alternative data-driven solution that preemptively gathers and
combines relevant data on three risk dimensions—Lack of Coping
Capacity, Vulnerability and Hazard & Exposure, according to the
INFORM Risk Index [
15
]—to provide a detailed risk assessment for
areas and communities in a country. This risk assessment supports a
faster identication of priority areas for humanitarian intervention
related to natural disasters. The output is an easy-to-understand
dashboard in which colors on a map visualize the risk or predicted
damage for a specic area. Currently, this toolbox is under develop-
ment for up to 10 countries (status January 2018) [
2
]. To be able to
develop this model for a country, a humanitarian data ecosystem
needs to be created in which data from dierent administrative
levels can be collected, collated and users are stimulated to make
use of the model.
As potential users of the humanitarian data ecosystem frame-
work, we asked seven interviewees involved in the development
of the Community Risk Assessment and Prioritization toolbox to
assess whether additional criteria could be identied that were
relevant specically for the humanitarian sector.
Secondly, to check for the practical relevance of the framework
criteria on a regional/country-specic level, interviews with eight
humanitarian data experts were held. These data experts repre-
sented ve dierent regions (Greater Horn of Africa
1
, Latin Amer-
ica and Caribbean, Malawi, the Philippines and the Southern Africa
Development Community
2
), to assess a variety of data landscapes
and to be able to distinguish between country-specic and generic
1
The Greater Horn of Africa comprises eight countries: Djibouti, Eritrea, Ethiopia,
Kenya, Somalia, South Sudan, Sudan and Uganda.
2
The Southern Africa Development Community (SADC) has 16 member states: Angola,
Botswana, Comoros, Democratic Republic of the Congo, Lesotho, Madagascar, Malawi,
answers. They all had knowledge on the broader data landscape
in their country, and experience with collecting data similar (or
identical) to the input data for the Community Risk Assessment
and Prioritization toolbox on low administrative levels, which is
necessary for the toolbox to generate accurate outcomes. In the
following section, we present the nal framework for developing a
data ecosystem for humanitarian purposes.
6 PHASE 3 - FRAMEWORK FOR
HUMANITARIAN DATA ECOSYSTEMS
The practical validation showed that the generic framework, which
was predominantly based on literature from the open data research
domain, needed to be adapted in several aspects to make it suitable
for the context of the humanitarian data ecosystem. Firstly, the
addition of a staged approach towards sets of criteria was proposed
by the developers of the Community Risk Assessment and Priori-
tization toolbox. Secondly, the importance of the (development of
the) governance part of the framework was recognized. And thirdly,
the empirical evidence as put forward by the expert interviews led
to an adaptation of the (classes of) governance criteria. We will
elaborate on these adaptations in the following paragraphs.
First, the developers of the Community Risk Assessment and
Prioritization toolbox indicated that the ‘data supply’ and ‘user
characteristics’ criteria both relate to the design and development
of the toolbox. In contrast, the ‘governance’ criteria relate to the
creation of an optimal context for the development of the toolbox.
Having practical experience with the development of the toolbox,
they evaluated the relevance of the ‘data supply’ and ‘user character-
istics’ criteria and conrmed that all related criteria were relevant
in the context of the humanitarian data ecosystem. However, they
also stressed that some of the elements will receive a stronger focus
in a later stage of the data ecosystem, stating that the humanitarian
data ecosystem around the Community Risk Assessment and Prior-
itization toolbox is still in a very early phase of development. In the
current stage, the focus is mainly on the data supply criteria. These
are directly related to data collection and collation, which is essen-
tial to be able to start developing the Community Risk Assessment
and Prioritization toolbox in the rst place. To illustrate this, the
criterion of data standards for example, is indeed ‘nice to have’ and
makes the data collection and analysis easier, but it is also possible
to get the data ecosystem running without data standards in place.
In the early stage of the humanitarian data ecosystem development,
also the user characteristics require extra attention. If the outcomes
of the toolbox do not match the user capabilities or information
needs of humanitarian decision-makers, the desired eect will not
be achieved.
The second result of the empirical validation was the labeling of
the governance part of the framework as important (being a major
challenge when developing a data ecosystem). This is indeed an
issue with the humanitarian data ecosystem around the Community
Risk Assessment and Prioritization toolbox: there are no, or very
few, governance structures in place yet that can stimulate the data
ecosystem development. Many fragmented organizations are in-
volved that are inclined to pursue their own objectives, while there
Mauritius, Mozambique, Namibia, Seychelles, South Africa, Swaziland, Tanzania, Zam-
bia and Zimbabwe.
is a lack of central governance to coordinate and align all these
individual activities. Berens et al. [
6
] refer to this situation as the
‘governance disparity’, which they describe as follows: “In humani-
tarian work, the multitude of independent organizations, government
actors, multinational initiatives, individuals, private sector recourse
providers, and digital platforms constitute a dynamic ecosystem with
no clear leader or dominant force” [26, p.1].
Thirdly, the interviews with eight humanitarian data experts
representing ve dierent regions, and one having global experi-
ence, yielded an overview of the main practical implications of the
theoretical governance criteria in the current humanitarian data
ecosystem. Assuming that a humanitarian data ecosystem spans
one country, and that the governance is taken care of within this
country, the data experts oered the following implications that
need to be included when setting up a humanitarian data ecosystem
within a country:
(1) Vision, communication and stimulation:
To create a
sense of responsibility and ownership, to have all ecosys-
tem stakeholders understand the value and importance of
data-sharing and to address the issue of fragmentation, it
is important that the party in the leadership role actively
involves all other relevant stakeholders in the establishment
of the data ecosystem.
(a) Division of roles and responsibilities:
In the specic
context of the Community Risk Assessment and Prioritiza-
tion toolbox, there are four dierent roles to be identied
and assigned, of which the rst two have a leading role
(and hence also relate to DG2 – Leadership and incen-
tivization in the framework):
The initiator/coordinator, which is the leading agency,
most likely a United Nations agency, in setting up the
humanitarian data ecosystem. In the toolbox context,
this party should initiate, coordinate and support the
development of the toolbox in a country, bring all rele-
vant actors together and incentivize them to facilitate
data-sharing and to use the toolbox.
The local lead, which is a national body that becomes
the local manager after the initiator/coordinator has set
out the initial data ecosystem, and is responsible for
locally promoting participation. This role is important
to ensure the sustainability of the ecosystem.
Data providers should be identied and are responsible
for sharing their data that is considered relevant.
Toolbox users, which are humanitarian or government
agencies who should use the toolbox (i.e. the data pro-
cessed into understandable information) and incorpo-
rate the outputs in their operations and decision-making.
(b) Feedback:
The possibility to provide feedback to data
providers is considered useful, but irrelevant in the cur-
rent phase of the humanitarian data ecosystem: collecting
relevant primary data and collating existing data is already
suciently challenging. The interviewees rather look for
alternatives than putting eort into the development of
feedback mechanisms. It is relevant, but in a later, more
developed stage of the humanitarian data ecosystem. The
same holds for transparency from data providers on how
they got their data—‘transparency’ is considered to be
overlapping with ‘feedback’.
(2) Leadership and incentivization:
The data experts sug-
gested to split up this element, as both aspects can be as-
sessed individually. ‘Leadership’ was already addressed with
the division of roles and ‘responsibilities’ (indicating an
overlap) and incentivization is considered very important,
and something that relates to all identied roles in the data
ecosystem. One of the most suggested approaches to incen-
tivize parties to participate in the data ecosystem was by
meeting the vision, ‘communication and stimulation’ crite-
rion: to involve stakeholders from the start.
(3) Participatory capacity:
In addition to the importance of
incentivization, the experts also labeled ‘participatory ca-
pacity’ as very important, and something that needs to be
addressed and built for every single role in the ecosystem.
This criterion is essential in the humanitarian data ecosys-
tem, as stakeholders generally lack technical and information
management capacities, especially at lower administrative
levels.
(4) Sustainability (nancing and value creation):
Sustain-
able nancing should be the responsibility of the initiator/-
coordinator of the data ecosystem. The criterion of ‘value
creation’ in the general framework is considered to be over-
lapping with incentivization as parties can be incentivized
to participate by creating value for them, hence the experts
consider ‘value creation’ as a criterion of ‘incentivization’.
(5) Trust and transparency:
The experts suggested to merge
transparency with ‘feedback’ as they relate to each other.
Trust between actors in the data ecosystem is indeed consid-
ered important. However, it is not a criterion for a successful
humanitarian data ecosystem, but rather something that fol-
lows from the creation of the vision, ‘communication and
stimulation’ criterion.
Based on the empirical validation, the governance part of the
framework was changed by adapting the (classication of) the
governance criteria. The changes are indicated by a bold border in
Figure 2.
By means of the empirical validation we adapted the generic
framework based on the criteria of developing ecosystems for open
data research to t with the context of the humanitarian sector. In
the following section, we discuss and reect on this deliverable and
oer future research topics.
7 DISCUSSION AND CONCLUSION
This study has taken a ‘data ecosystem’ approach to create a frame-
work to guide the facilitation, stimulation and coordination of data
collection, collation and data sharing between humanitarian actors
to increase the level of data preparedness in a country. By doing
so, this study is the rst academic study to address the ‘humani-
tarian data ecosystem’. A framework of criteria for the successful
development of a data ecosystem has been developed based on a
systematic literature review and was validated with a panel of open
data experts. Subsequently, this framework has been tested on its
application and relevance in the humanitarian sector by means of
a case study. The need to test this stemmed from the fact that the
Figure 2: Governance criteria after empirical validation
ecosystem approach has so far mainly been applied in the research
domain of open data. However, the issue in the humanitarian sector
is that much data is not shared openly, and sometimes not even
stored digitally. The question to be answered was therefore: should
these same criteria be considered when establishing a data ecosys-
tem in the complex, fragmented humanitarian context around the
Community Risk Assessment and Prioritization toolbox developed
by the Netherlands Red Cross?
We found that rather than classifying the humanitarian data
ecosystem as a distinct type of data ecosystem, it is a data ecosystem
that is in an emerging phase of development whereby it links to
other "sub-data ecosystems or in some cases data collaboratives",
that collect, handle and use data on specic risk indicators (such as
in relation to the health sector) - with a relatively long retention
period - or crisis data with a short retention period. To stimulate its
development, the special point of attention in the humanitarian data
ecosystem appeared to be the governance part of the framework.
Therefore, eight respondents representing ve dierent regions
with varying data landscapes were asked about how they perceived
the relevance and application of the dierent governance criteria
in their region. It was hereby assumed that a humanitarian data
ecosystem spans one country, and that the governance is taken
care of within this country. The interviewees also specied the
practical implications of the criteria for the humanitarian data
ecosystem specically ‘Value creation’ and ‘trust’ were found to be
overlapping with other criteria and hence omitted, but all the other
criteria were considered very relevant. A remark must be made for
‘transparency and feedback’ though; according to the interviewees
these only become relevant in a later data ecosystem stadium, due
to the current focus on enhancing, facilitating and speeding-up
data-sharing practices.
By means of the interviews, it was also found that many of the
governance criteria closely interrelate with ‘data supply’ and ‘user
characteristics’ criteria: they are all interdependent. The division
of roles and responsibilities for example, should also include the
identication of data providers – relating to the data supply part
of the framework, and toolbox users – relating to the user charac-
teristics part of the framework. Moreover, the ‘leadership’ is also a
part of the division of roles and ‘responsibilities’, and likewise ‘in-
centivization’ and ‘participatory capacity’ hold for every identied
role.
Approaching data-sharing in the humanitarian sector with a
data ecosystem perspective provides insight into the elements to
be addressed and helps to provide structure. However, the interre-
lations between the criteria are a lot more complex than visualized
in Figure 1, which indicates that the humanitarian data ecosystem
might be too complex to capture in a static framework. Additionally,
some of the criteria are considered to be of secondary importance:
they only become relevant in a late stage of the data ecosystem.
Susha, Janssen & Verhulst [
28
] try to capture the more dynamic
elements through introducing the concept of Data Collaboratives, a
cross-sector (and public-private) collaboration initiative aiming at
data collection, sharing, or processing for the purpose of addressing
a societal challenge.
To summarize, the humanitarian data ecosystem is emerging
and currently in its initial development phase. There is a strong
focus on ‘data supply’ and on – though less but still present – ‘user
characteristics’, but the governance part requires attention to create
an optimal context to facilitate meeting the data supply and user
characteristics criteria.
Future research should extend this study by also testing the the-
oretical framework of data ecosystem success criteria to a set of
other case studies, as an extra validation. This would both provide
insight into the generic applicability of the framework, and in the
specicity of the ndings for the humanitarian data ecosystem.
Other case studies can also provide insight into the evolvement
of the maturity of other data ecosystems, which is something the
immature humanitarian data ecosystem can learn from. Moreover,
future research should address the identication and establishment
of the IT infrastructures, including geospatial data-sharing plat-
forms, required to support the development of the humanitarian
data ecosystem. The architectural design of these infrastructures
should match the local context, and the capabilities of the stakehold-
ers in the data ecosystem. We also recommend future research into
potential undesired consequences of setting up a humanitarian data
ecosystem. The selection of actors to be included or excluded from
the data ecosystem can lead to shifts in tasks and, consequently,
power balances amongst stakeholders, which can lead to power
abuse. Additionally, as the data will also include personal data, com-
pliance with personal data protection laws and regulation need to
be carefully assessed. In general, unethical use of data (be it per-
sonal or other types of data) needs to be carefully assessed. Lastly,
the application in practice of the developed governance structure
should be evaluated. For this purpose, actual use of the framework
in a diversity of countries can yield even more empirical evidence
for either conrmation or renement of (the components and their
criteria in) the framework. This empirical approach will also enable
the inclusion of the viewpoints of other actors than the data experts
to which our study was limited.
ACKNOWLEDGMENTS
We would like to thank the contributors to this study: all intervie-
wees, Jannis Visser for developing the Community Risk Assessment
and Prioritization toolbox, and the Princess Margriet Fund for -
nancing the initial toolbox development.
REFERENCES
[1]
510 Global. 2016. Personalised Humanitarian Aid Through Big Data: an initiative
of the Dutch Red Cross. (2016). http://510.global (accessed: March 21, 2017).
[2]
510 Global. 2017. Community Risk Assessment and Prioritization toolbox. (2017).
https://dashboard.510.global/ (accessed: March 21, 2017).
[3]
Judie Attard, Fabrizio Orlandi, and Sören Auer. 2016. Data Value Networks:
Enabling a New Data Ecosystem. In Proc. IEEE/WIC/ACM International Conference
on Web Intelligence (WI). 453–456.
[4]
Florian Barthélemy. 2016. " The Belgian open data ecosystem and innovation
through open data. (2016).
[5]
A Benini. 2015. The use of Data Envelopment Analysis to calculate priority
scores in needs assessments. Active Learning Network for Accountability and
Performance in Humanitarian Action (ALNAP), London, UK (2015).
[6]
J Berens, N Raymond, G Shomshon, S Verhulst, and L Bernholz. 2015.
The Humanitarian Data Ecosystem: the Case for Collective Responsibil-
ity. (2015). https://pacscenter.stanford.edu/wpcontent/uploads/2017/08/
humanitarian_data_ecosystem.pdf (accessed: September 19, 2017).
[7]
Tina Comes. 2016. Cognitive biases in humanitarian sensemaking and decision-
making lessons from eld research. In Proc. IEEE International Multi-Disciplinary
Conference on Cognitive Methods in Situation Awareness and Decision Support
(CogSIMA). 56–62.
[8]
Tim Davies. 2011. Open Data: infrastructures and ecosystems. Open Data Research
(2011).
[9]
Sharon S Dawes, Lyudmila Vidiasova, and Olga Parkhimovich. 2016. Planning and
designing open government data programs: An ecosystem approach. Government
Information Quarterly 33, 1 (2016), 15–27.
[10]
Teresa M Harrison, Theresa A Pardo, and Meghan Cook. 2012. Creating open
government ecosystems: A research and development agenda. Future Internet 4,
4 (2012), 900–928.
[11]
Maximilian Heimstädt, Fredric Saunderson, and Tom Heath. 2014. Conceptualiz-
ing Open Data ecosystems: A timeline analysis of Open Data development in the
UK. In Proc. Conference for E-Democracy and Open Governement. 245.
[12]
Dorothea Hilhorst. 2013. Disaster, conict and society in crises: everyday politics
of crisis response. Routledge.
[13]
IEAG, United Nations. 2014. A World that Counts–Mobilising the Data Revolution
for Sustainable Development. (2014). http://www.undatarevolution.org/report
(accessed: January 24, 2018).
[14]
Anne Immonen, Marko Palviainen, and Eila Ovaska. 2014. Requirements of an
open data based business ecosystem. IEEE access 2 (2014), 88–103.
[15]
INFORM. 2016. Methodology. (2016). http://www.inform-index.org/InDepth/
Methodology (accessed: June 24, 2017).
[16]
Thorhildur Jetzek. 2017. Innovation in the Open Data Ecosystem: Exploring the
Role of Real Options Thinking and Multi-sided Platforms for Sustainable Value
Generation through Open Data. In Analytics, Innovation, and Excellence-Driven
Enterprise Sustainability. Springer, 137–168.
[17]
Deirdre Lee. 2014. Building an open data ecosystem: an Irish experience. In Proc.
International Conference on Theory and Practice of Electronic Governance. ACM,
351–360.
[18]
Cathy Macharis and Joep Crompvoets. 2014. A stakeholder-based assessment
framework applied to evaluate development scenarios for the spatial data in-
frastructure for Flanders. Computers, Environment and Urban Systems 46 (2014),
45–56.
[19]
Asmat Monaghan and Mark Lycett. 2013. Big data and humanitarian supply
networks: Can Big Data give voice to the voiceless?. In Proc. IEEE Global Human-
itarian Technology Conference (GHTC). 432–437.
[20]
S O’Brien. 2017. This is how we build a stronger, data-driven hu-
manitarian sector. (2017). https://www.weforum.org/agenda/2017/01/
this-is- how-we- build- a-stronger- data-driver-humanitarian%20sector (accessed:
July 3, 2017).
[21]
MT Ortuño, P Cristóbal, JM Ferrer, FJ Martín-Campo, S Muñoz, G Tirado, and B
Vitoriano. 2013. Decision aid models and systems for humanitarian logistics. A
survey. In Decision aid models for disaster management and emergencies. Springer,
17–44.
[22]
Mark A Parsons, Øystein Godøy, Ellsworth LeDrew, Taco F De Bruin, Bruno
Danis, Scott Tomlinson, and David Carlson. 2011. A conceptual framework for
managing very diverse data for complex, interdisciplinary science. Journal of
Information Science 37, 6 (2011), 555–569.
[23]
Rufus Pollock. 2011. Building the (open) data ecosystem. Open Knowledge
Foundation Blog 31 (2011).
[24] Diego Ponte. 2015. Enabling an Open Data Ecosystem.. In ECIS.
[25]
N Raymond and Z Al Achkar. 2016. Data preparedness: connecting data, decision-
making and humanitarian response. (2016). http://hhi.harvard.edu/sites/default/
Harvard Humanitarian Initiative: signal program on human security and tech-
nology - standards and ethics series 01, (accessed: July 4, 2017).
[26]
N Raymond, Z Al Achkar, S Verhulst, and J Berens. 2016. Building data respon-
sibility into humanitarian action. (2016). http://www.unocha.org/sites/dms/
Documents/TB18_Data%20Responsibility_Online.pdf OCHA Policy and Studies
Series 018, (accessed: July 4, 2017).
[27]
J Sabou. 2016. The Humanitarian “Data” Ecosystem: Intersection of Innovation
& Disaster Relief. (2016). http://www.leidensafetyandsecurityblog.nl/articles/
the-humanitarian- data-ecosystem- intersection- of-innovation- disaster- relief
(accessed: July 3, 2017).
[28]
Iryna Susha, Marijn Janssen, and Stefaan Verhulst. 2017. Data collaboratives as
a new frontier of cross-sector partnerships in the age of open data: taxonomy
development. (2017).
[29]
Bartel Van de Walle and Tina Comes. 2014. Risk accelerators in disasters: Insights
from the Typhoon Haiyan Response on Humanitarian Information Management
and Decision Support. In Proc. International Conference on Advanced Information
Systems Engineering. Springer, 12–23.
[30]
Marc van den Homberg, Jannis Visser, and Maarten van der Veen. 2017. Unpack-
ing Data Preparedness from a humanitarian decision making perspective: toward
an assessment framework at subnational level. In Proc. Information Systems for
Crisis Response and Management (ISCRAM) Conference.
[31]
Francois Van Schalkwyk, Michelle Willmers, and Maurice McNaughton. 2016.
Viscous open data: The roles of intermediaries in an open data ecosystem. Infor-
mation Technology for Development 22, sup1 (2016), 68–83.
[32]
Frederika Welle Donker and Bastiaan van Loenen. 2017. How to assess the
success of the open data ecosystem? International Journal of Digital Earth 10, 3
(2017), 284–306.
[33]
K Whipkey and A Verity. 2015. Guidance for incorporating Big Data into
Humanitarian Operations. (2015). http://digitalhumanitarians.com/resource/
incorporating-big- data-humanitarian-operations (accessed: April 5, 2017).
[34]
Martin Wiener, Friedrich T Sommer, Zachary G Ives, Russell A Poldrack, and
Brian Litt. 2016. Enabling an open data ecosystem for the Neurosciences. Neuron
92, 3 (2016), 617–621.
[35]
Anneke Zuiderwijk, Marijn Janssen, and Chris Davis. 2014. Innovation with
open data: Essential elements of open data ecosystems. Information Polity 19, 1,
2 (2014), 17–33.
... Numerous business, research, and industry communities study the big data field [19,20,68]. Some definitions stay relevant to specific domains like humanitarian [32,40] and personal data ecosystems [61]. These studies have a narrow perspective and focus on a certain notion with partial details [6,31,97]. ...
... The first definition explains the humanitarian data ecosystem as a network of humanitarian actors, governments, private sector organizations, and affected communities in which they interact with each other to produce, collect and analyze digital data about vulnerable populations [32,40]. Another definition describes the data ecosystem as a socio-technical complex network in which actors like organizations and individuals interact, collaborate to exchange and use data as the primary source to foster innovation and support new businesses [67]. ...
... "Data value creation" is about mining value from big data through its extensive use. We found a list of areas of data value creation from the literature including data-driven administration [95], public service delivery, data economy [22], innovation [22,35,68], new businesses [61,99], data economy [22], transparency [22,26,49,54,90], public policies and strategies alignments [39,88], open government [54], secure exchange of data [32], people's personal data management [29], government performance [90], democratic governance and political participation, trust in government, data reuse, and integration of public and private data [22]. ...
Article
The public sector, private firms, business community, and civil society are generating data that are high in volume, veracity, and velocity and come from a diversity of sources. This type of data is today known as big data. Public administrations pursue big data as “new oil” and implement data-centric policies to collect, generate, process, share, exploit, and protect data for promoting good governance, transparency, innovative digital services, and citizens’ engagement in public policy. All of the above constitute the Government Big Data Ecosystem (GBDE). Despite the great interest in this ecosystem, there is a lack of clear definitions, the various important types of government data remain vague, the different actors and their roles are not well defined, while the impact in key public administration sectors is not yet deeply understood and assessed. Such research and literature gaps impose a crucial obstacle for a better understanding of the prospects and nascent issues in exploiting GBDE. With this study, we aim to start filling the above-mentioned gaps by organizing our findings from an extended Systematic Literature Review into a framework to organise and address the above-mentioned challenges. Our goal is to contribute in this fast-evolving area by bringing some clarity and establishing common understanding around key elements of the emerging GBDE.
... In the face of these challenges, the humanitarian response and its actors are demonstrating their resilience and their capacity always to renew and progress to help more, help further, and to help better with financial means that have certainly improved since the Grand Bargain in 2016 but are still too uneven and insufficient (Development Initiatives, 2020). The rapid increase in data availability leads to a shift towards a more decision-making approach to humanitarian action, providing humanitarian actors with the means to become more efficient and costeffective (Haak et al., 2018). Coupled with increasingly effective algorithms outperforming some human capacities and advanced calculation storage (such as graphics processing units, GPUs), the humanitarian information environment might face a significant renewal with Artificial Intelligence (United Nations, 2019). ...
... Coupled with increasingly effective algorithms outperforming some human capacities and advanced calculation storage (such as graphics processing units, GPUs), the humanitarian information environment might face a significant renewal with Artificial Intelligence (United Nations, 2019). While Artificial Intelligence (AI) development and research have focused mainly on commercial interests in recent years (Fosso Wamba et al., 2021), AI affects the way all organization structures deal with information and, therefore, how they organize themselves to make a decision (Haak et al., 2018). ...
... It is now ten years since it is common to hear set phrases like 'Data is the new oil' or to compare data as a new economic input into an organization as is capital or labor (The Economist 2010). Coupled with a massive increase in the availability of data enabled by the digitization of our world (Haak et al., 2018), algorithms and Artificial Intelligence may indeed have the potential to unlock new sources of economic and social value, leading first and foremost to a growing interest and investment in the data industry in the private sector (Bell et al., 2020). We therefore understand that before being able to do machine learning or AI, any organization must be interested in structuring its data, studying how to organize it, and establishing a consistent framework. ...
Thesis
Full-text available
During a humanitarian crisis the major challenge of the decision-making process is to access to as much insightful information as possible. The fast development of artificial intelligence and machine learning over the last decades is revolutionizing how information is generated from 'Big Data' in the private and the public sectors. The main objective of this paper is to study how artificial intelligence and machine learning can transform decision-making in humanitarian disaster management and to assess how impactful this transformation can be. In that respect, this paper aims at providing managers with practical knowledge and use cases of how AI applications can support them to make decisions to respond to crises. The approach taken here is to use a high-end view of AI and its dimensions according to the latest research in machine learning and deep learning. A careful analysis on data ecosystem and the managerial challenges of adopting AI in humanitarian culture is needed prior to AI implementation. Our paper shows that artificial intelligence and machine learning have both a positive impact on the four phases of the decision-making process of disaster management (mitigation, preparedness, response and recovery) but also will have a high impact in the future that will grow the more the sector adopts these technologies.
... In the development of EWS, many countries and organizations have shifted towards using more data-driven and high-tech tools as this would foster more efficient, effective, and evidence-based decision-making, and can increase the lead time to prepare and anticipate to floods (Budimir et al., 2020;Haak et al., 2018). Both in their design and function, EWS increasingly depend on computing power (Budimir et al., 2020). ...
... Decision-making based on up-to-date, validated, and reliable information is often seen as a main challenge in working with EWS, due to stress, time pressure, and uncertain information (Comes, 2016;Haak et al., 2018). Several studies have suggested that using data can help shift towards more efficient, more effective, and more evidence-based decision-making in this context (Benini, 2015;Haak et al., 2018). ...
... Decision-making based on up-to-date, validated, and reliable information is often seen as a main challenge in working with EWS, due to stress, time pressure, and uncertain information (Comes, 2016;Haak et al., 2018). Several studies have suggested that using data can help shift towards more efficient, more effective, and more evidence-based decision-making in this context (Benini, 2015;Haak et al., 2018). Therefore, small and big data are increasingly combined to improve the functioning of flood EWS. ...
Conference Paper
Full-text available
Increasingly, data-driven instruments are used in disaster risk reduction to foster more efficient, effective, and evidence-based decision-making. This data revolution brings along opportunities and challenges, which are sometimes related to the data itself, but more often seem related to the environment in which the data is put to use. To provide insight into such an emerging data ecosystem, this paper uses a qualitative case study to assess the use of data in flood early warning systems (EWS) in Nepal. In response to the research question ‘How does the data ecosystem impact the opportunities and challenges regarding data use in flood early warning systems in Nepal?’, this paper discusses the importance of considering the broader context instead of regarding data as an entity unto itself. It shows how actors, policies and other contextual factors impact the effectiveness of data use by either presenting opportunities, like the establishment of a national disaster data repository, or challenges, like inadequate human resources for working with data.
... Numerous business, research, and industry communities study the big data field [76][77][78]. For example, in the case of definitions of BDE, some definitions stay relevant to specific domains like humanitarian [79][80][81], and personal data ecosystems [67]. Such studies have a narrow perspective, focus on a specific notion with partial details, and describe BDE definitions and other related terminologies that vary considerably [81][82][83]. ...
... However, PAs do not publish various data sets due to certain data traits, like data containing personal or sensitive information [175]. PAs intend to publish government data for all to promote transparency, accountability, value creation, i.e., better governance, and to enhance the quality of life of the citizens [67,79,175]. The data publish phase is highly essential for the open government domain. ...
... • Ensure to take appropriate measures that enable individuals to control whom to share data and how much the owner is eager to share [67,97,114,142]. • Data providers focus on maintaining a balance between data availability and data redundancy when publishing data through various formats [79,93]. • Consider data sharing granularity and data transmission in addition to authorization of data while sharing private data. ...
Article
Full-text available
The public sector, private firms, business community, and civil society are generating data that is high in volume, veracity, velocity and comes from a diversity of sources. This kind of data is known as big data. Public Administrations (PAs) pursue big data as “new oil” and implement data-centric policies to transform data into knowledge, to promote good governance, transparency, innovative digital services, and citizens’ engagement in public policy. From the above, the Government Big Data Ecosystem (GBDE) emerges. Managing big data throughout its lifecycle becomes a challenging task for governmental organizations. Despite the vast interest in this ecosystem, appropriate big data management is still a challenge. This study intends to fill the above-mentioned gap by proposing a data lifecycle framework for data-driven governments. Through a Systematic Literature Review, we identified and analysed 76 data lifecycles models to propose a data lifecycle framework for data-driven governments (DaliF). In this way, we contribute to the ongoing discussion around big data management, which attracts researchers’ and practitioners’ interest.
... We added incentive to share data as this was not a distinguishing characteristic for a data collaborative, but it is among actors in a data ecosystem. Haak et al. [29] developed a framework of criteria for a successful data ecosystem specifically for humanitarian purposes, including data supply, user characteristics and governance criteria. The framework of Haak et al. describes in more detail additional elements of a data ecosystem not explicitly covered in the work of Susha et al., such as data governance and data infrastructures. ...
... An Integrated Data Ecosystem Framework Figure 1 and Table 1 present an integrated framework to characterize data ecosystems which combines the relevant existing frameworks [14,29,30] and further details and elaborates on additional elements. The framework is structured around five dimensions: actors, data supply, data infrastructure, data demand and data ecosystem governance, whereby each dimension has different indicators. ...
... The data ecosystem governance comprises the framework of policies, processes and instruments to realize common goals in the interaction between entities [29]. The different elements we selected were participatory capacity, continuity of collaboration between users and suppliers, communication, incentive to share and use data, user selection and collaboration among data users. ...
Article
Full-text available
Reporting on the Sustainable Development Goals (SDGs) is complex given the wide variety of governmental and NGO actors involved in development projects as well as the increased number of targets and indicators. However, data on the wide variety of indicators must be collected regularly, in a robust manner, comparable across but also within countries and at different administrative and disaggregated levels for adequate decision making to take place. Traditional census and household survey data is not enough. The increase in Small and Big Data streams have the potential to complement official statistics. The purpose of this research is to develop and evaluate a framework to characterize a data ecosystem in a developing country in its totality and to show how this can be used to identify data, outside the official statistics realm, that enriches the reporting on SDG indicators. Our method consisted of a literature study and an interpretative case study (two workshops with 60 and 35 participants and including two questionnaires, over 20 consultations and desk research). We focused on SDG 6.1.1. (Proportion of population using safely managed drinking water services) in rural Malawi. We propose a framework with five dimensions (actors, data supply, data infrastructure, data demand and data ecosystem governance). Results showed that many governmental and NGO actors are involved in water supply projects with different funding sources and little overall governance. There is a large variety of geospatial data sharing platforms and online accessible information management systems with however a low adoption due to limited internet connectivity and low data literacy. Lots of data is still not open. All this results in an immature data ecosystem. The characterization of the data ecosystem using the framework proves useful as it unveils gaps in data at geographical level and in terms of dimensionality (attributes per water point) as well as collaboration gaps. The data supply dimension of the framework allows identification of those datasets that have the right quality and lowest cost of data extraction to enrich official statistics. Overall, our analysis of the Malawian case study illustrated the complexities involved in achieving self-regulation through interaction, feedback and networked relationships. Additional complexities, typical for developing countries, include fragmentation, divide between governmental and non-governmental data activities, complex funding relationships and a data poor context.
... e EHRQA technique, which is part of the leadership model, is being implemented progressively in order to alter the old talent management strategy. Comparing the leadership model to the typical EHRQA approach, the leadership model places a greater emphasis on the applicability of employees to the organization and pays more attention to the workability and performance of employees while at work [7][8][9][10][11][12]. Evaluating employees' initiative, creativity capacity, and cooperation ability, among other traits, allows them to maximize their own initiative and maximize the value of their own abilities, thus enabling the firm to enter a new stage of development [13,14]. ...
Article
Full-text available
With the advancement of society and economy, the market competition among various businesses has become increasingly fierce. Nowadays, if businesses want to grow in the face of adversity, they must move forward boldly and make use of abundant human resources to fuel their growth. Human resource management is becoming increasingly important. As a result, this paper develops an enterprise human resource quality assessment system based on the WICS leadership model. The differences between the WICS model and the traditional management model are first compared in this paper. The requirements of the WICS model in human resource management are then described. Furthermore, this paper proposes a human resource evaluation algorithm that combines data-driven and WICS models to address the current human resource cost evaluation algorithm's low accuracy and poor effect. The simulation results show that the proposed algorithm can reflect changing human resource cost characteristics, improve human resource cost evaluation results, and obtain better results than other human resource cost evaluation models and has a wide range of applications.
... This is mostly due to the fact that it is often not clear what incentives and ultimately what added value data providers have when offering their data to other actors in an ecosystem (Badewitz et al. 2020;van den Homberg and Susha 2018). For this reason, fair and reasonable incentive and revenue distribution mechanisms are important for reliable cooperation and sustainable ecosystem development (Cappiello et al. 2019;Haak et al. 2018;Shen et al. 2020). ...
Conference Paper
Full-text available
In the increasingly connected business world, economic value is created less and less by one company alone but rather through the combination and enrichment of data by various actors in so-called data ecosystems. However, one of the main obstacles to why actors are currently not motivated to engage in data ecosystems is that they are often not aware of the actual benefits of cross-organisational data sharing. This is partly because it is in many cases unclear what the incentives and ultimately the added value are for data providers when they share their data with others. To address this research gap we develop a taxonomy of incentive mechanisms for data sharing in data ecosystems which is based on a structured literature review. The resulting taxonomy consists of key dimensions and characteristics of incentive mechanisms for data sharing in data ecosystems and contributes to a better scientific understanding of these concepts.
... Based on the works of [48] and [49], [50] developed a framework to characterize data ecosystems based on five dimensions. This framework is, however, focused on the description of data ecosystems in developing countries. ...
Conference Paper
Full-text available
In the increasingly interconnected business world, economic value is less and less created by one company alone but rather through the combination and enrichment of data by various actors in so-called data ecosystems. The research field around data ecosystems is, however, still in its infancy. With this study, we want to address this issue and contribute to a deeper understanding of data ecosystems. Therefore, we develop a taxonomy for data ecosystems which is grounded both theoretically through the linkage to the scientific knowledge base and empirically through the analyses of data ecosystem use cases. The resulting taxonomy consists of key dimensions and characteristics of data ecosystems and contributes to a better scientific understanding of this concept. Practitioners can use the taxonomy as an instrument to further understand, design and manage the data ecosystems their organizations are involved in.
... There is an emerging literature on how to characterize and govern data ecosystems. Haak et al. (2018) developed a framework of criteria for a successful data ecosystem specifically for humanitarian purposes, including data supply, user characteristics and governance criteria. Instead of the overarching data ecosystem approach, one can also segment a data ecosystem into data collaboratives, i.e. cross-sector (and public-private) collaboration initiatives aiming at data collection, sharing, or processing for the purpose of addressing a societal challenge (Susha et al., 2017). ...
Conference Paper
Full-text available
The importance of data as the key resource in start-ups and traditional businesses rises steadily. Thus, traditional organizations need to adapt accordingly and develop adequate strategies to incorporate data into their value creation process. Creating value from data often requires collaboration between various actors with different requirements and individual objectives, in short, in an ecosystem. However, as of now, there are few studies investigating data ecosystems in-depth. Thus, we address that issue and report on a Design Science Research study developing a morphology of data ecosystems based both on a structured literature review and two explorative case studies conducted in traditional organizations. Therefore, we generate new knowledge through first-hand insights gained via use case studies over 18 months with practitioners and support these results with existing expertise in the scientific literature. We created a morphology to reveal key characteristics and elements of data ecosystems to develop a tool for companies and scientists alike to acquire a better understanding and perception of data ecosystems .
Conference Paper
Full-text available
All too often the collection as well as analysis of data for humanitarian response only starts once a disaster hits. This paper proposes a framework to assess Data Preparedness on five dimensions: Data Sets, Data Services and Tooling, Data Governance, Data Literacy, and Networked Organizations for Data. We demonstrate for one dimension, i.e. Data Sets, how it can be quantified. First step is to determine which Data Sets must be collected before a disaster strikes so that as many as possible decision-makers’ information needs are covered. Subsequently, a Data Sets Preparedness Index can be calculated based on Completeness, Recency and Accuracy & Reliability. We tested the index for Malawi and The Philippines and show how it can be used to direct data collection and determine when data analysis for e.g. predicting severity becomes meaningful. The index can be modified for reporting on global policies such as the Sustainable Development Goals. http://idl.iscram.org/files/marcvandenhomberg/2017/1995_MarcvandenHomberg_etal2017.pdf
Chapter
Full-text available
While open data as a phenomenon is rapidly growing up, innovation through open data is still less than expected. Research has shown that in spite of emerging new businesses models, private sector stakeholders are struggling to generate monetary income from open data. This is worrying as open data initiatives might not be sustained if there no evidence of value generation through external use of the data. We suggest that insights from two established theories, real options theory and theory of two-sided markets, might help us create a more coherent picture of the complex relationships between innovation and value generation in the open data ecosystem, and even resolve what we call the open data value paradox. We propose that governments, which openly publish data, are providing private sector stakeholders with the equivalent of a real option. By conceptualizing the uncertain or serendipitous value of open government data as option value, we might be able to stimulate activity and investment in the open data ecosystem. Moreover, we propose that by utilizing two-sided markets type of business models, private companies can use the data as a resource to provide free information and by capitalizing on the resulting positive network externalities, generate monetary income as well. Finally, we propose that governments should provide the necessary nourishment to this ecosystem in order to stimulate the generation of sustainable value.
Article
Full-text available
Open data are currently a hot topic and are associated with realising ambitions such as a more transparent and efficient government, solving societal problems, and increasing economic value. To describe and monitor the state of open data in countries and organisations, several open data assessment frameworks were developed. Despite high scores in these assessment frameworks, the actual (re)use of open government data (OGD) fails to live up to its expectations. Our review of existing open data assessment frameworks reveals that these only cover parts of the open data ecosystem. We have developed a framework, which assesses open data supply, open data governance, and open data user characteristics holistically. This holistic open data framework assesses the maturity of the open data ecosystem and proves to be a useful tool to indicate which aspects of the open data ecosystem are successful and which aspects require attention. Our initial assessment in the Netherlands indicates that the traditional geographical data perform significantly better than non-geographical data, such as healthcare data. Therefore, open geographical data policies in the Netherlands may provide useful cues for other OGD strategies.
Article
Full-text available
The open government data (OGD) movement has rapidly expanded worldwide with high expectations for substantial benefits to society. However, recent research has identified considerable social and technical barriers that stand in the way of achieving these benefits. This paper uses sociotechnical systems theory and a review of open data research and practice guidelines to develop a preliminary ecosystem model for planning and designing OGD programs. Findings from two empirical case studies in New York and St. Petersburg, Russia produced an improved general model that addresses three questions: How can a given government's open data program stimulate and support an ecosystem of data producers, innovators, and users? In what ways and for whom do these the ecosystems produce benefits? Can an ecosystem approach help governments design effective open government data programs in diverse cultures and settings? The general model addresses policy and strategy, data publication and use, feedback and communication, benefit generation, and advocacy and interaction among stakeholders. We conclude that an ecosystem approach to planning and design can be widely used to assess existing conditions and to consider policies, strategies, and relationships that address realistic barriers and stimulate desired benefits.
Article
As the pace and complexity of neuroscience data grow, an open data ecosystem must develop and grow with it to allow neuroscientists the ability to reach for new heights of discovery. However, the problems and complexities of neuroscience data sharing must first be addressed. Among the challenges facing data sharing in neuroscience, the problem of incentives, discoverability, and sustainability may be the most pressing. We here describe these problems and provide potential future solutions to help cultivate an ecosystem for data sharing.
Article
Data.gov celebrated its fifth birthday in May 2014. Data.gov.uk will turn five in January 2015. According to datacatalogs.org, there are more than 380 Open Data portals around the world, with new portals still appearing all the time. Some Open Data practitioners cried 'go ugly, go early', others 'raw data now', while others lobbied for CSV and RDF. With Open Data initiatives maturing, what can countries such as Ireland, who are embarking on their national Open Data journey, learn from the experience of these instigators? What are best practices for selecting datasets, publishing data and supporting data reuse? In this paper, I present the experience of defining an Irish Open Data strategy, based on an ecosystem structure incorporating not only technical, but social, political, economic and legislative aspects. Each element of the ecosystem represents a key component of a sustainable Open Data initiative. Through an extensive literature review of international best practice and liaising with stakeholders, the ecosystem elements evolved, along with common principles and challenges. Finally, the outcome of this study is a concrete roadmap to progress the vision of Open Data in Ireland.
Article
Open data ecosystems are expected to bring many advantages, such as stimulating citizen participation and innovation. However, scant attention has been given to what constitutes an open data ecosystem. The objective of this paper is to provide an overview of essential elements of open data ecosystems for enabling easy publication and use of open data. To achieve this objective, the literature has been reviewed and a scenario about the publication and use of open data has been analyzed. It was found that various applications, tools and portals are available which together can form an ecosystem. The best functionalities of this ecosystem can be selected and utilized by open data providers and users. To create an open data ecosystem at least four key elements should be captured, namely, 1) releasing and publishing open data on the internet, 2) searching, finding, evaluating and viewing data and their related licenses, 3) cleansing, analyzing, enriching, combining, linking and visualizing data and 4) interpreting and discussing data and providing feedback to the data provider and other stakeholders. Furthermore, to integrate the ecosystem elements and to let them act as an integrated whole, there should be three additional elements 5) user pathways showing directions for how open data can be used, 6) a quality management system and 7) different types of metadata to be able to connect the elements.