Content uploaded by Ossi Ylijoki
Author content
All content in this area was uploaded by Ossi Ylijoki on Mar 09, 2017
Content may be subject to copyright.
1 (16)
Conceptualizing Big Data:
Analysis of Case Studies
OSSI YLIJOKIa* AND JARI PORRASa
a School of Business and Management, Lappeenranta University of Technology, Finland
*Corresponding author, address: Puistokatu 1C38, FI-15100 Lahti, Finland;
e-mail: ossi.ylijoki@phnet.fi; tel. +358442387442
ABSTRACT. Digitization and the related datafication produce huge amounts of data.
Organizations have started to exploit these new data in order to gain benefits. Exploring this
”big data jungle” is a new area for both scholars and practitioners, and the experiences of early
adopters are valuable. This paper analyses big data use cases described in the academic
literature by using computerized content analysis methods. Based on the analysis results, we
have conceptualized themes and guidelines of big data in the context of an organization, thus
contributing to the emerging research of big data. In addition to the realized benefits, the case
studies reveal issues regarding technology, skills, organizational culture, and decision-making
processes. The paper also points out several new research avenues, acts as a reference
collection to big data case studies found in academic sources, and bridges theory and practice
by pointing out several topics that practitioners should consider.
Keywords: big data, case study, content analysis, digital transformation, digitization
1. Introduction
Today, new digital technologies produce vast amounts of various types of data (Gantz and
Reinsel, 2011), often referred to as big data. From the point of view of technology, big data
are different from traditional transaction data, requiring new data management and analysis
technologies (Laney, 2001). More importantly, several sources, including (Davenport, 2014;
Manyika et al., 2011; Mayer-Schönberger and Cukier, 2013) claim that big data have
potentially huge effects on many industries. Technology and data drives change, and as e.g.
(Dehning et al., 2003; Sainio, 2005) suggest, companies must link their strategy with
technology. The business environment is changing. However, it is difficult to forecast the
impacts at the micro level, as digitization and data deluge are a new, emerging phenomenon.
The effects of this phenomenon are different for each company. As an example, self-
driving cars
1
, which will invade the markets in the future, will have significant effects on
various firms, like car dealers and insurance companies. However, the potential and the
challenges that a car dealer faces will differ significantly from those of an insurance company.
Realizing the potential implies that this new, data-driven paradigm will affect companies’
strategies and business models heavily. Several excellent pieces of work exist on business
transformation. Venkatraman (1994) builds a framework that helps understand the effects of
the transformation. Christensen (2013) explains clearly how incumbent companies fail
1
E.g. Google: http://googleblog.blogspot.fi/2015/05/self-driving-vehicle-prototypes-on-road.html,
Nissan: http://abcnews.go.com/Technology/nissan-driving-car-ready-2020-ceo/story?id=31120512,
or Volvo: http://www.wired.com/2015/02/volvo-will-test-self-driving-cars-real-customers-2017/
2 (16)
constantly in utilizing new, disruptive technologies. Sainio (2005) shows that companies are
often well aware of new, emerging technologies, but neglect linking the technologies with their
strategies.
There are some trailblazers, Google and Amazon being the most obvious examples, which
have built their business models around data. These kind of examples, as well as some previous
studies, e.g. (Dehning et al., 2003; McAfee et al., 2012; Porter and Millar, 1985) indicate that
companies utilizing data heavily gain competitive advantage over their less data-driven rivals.
However, the data-driven approach is still a new paradigm for most organizations (Shen and
Varvel, 2013). In addition, established companies have their own history, processes, and
capabilities. They just cannot turn their existing structures and business models upside down
at once. The transformation takes time. When established firms start to explore the possibilities
of big data, they can learn from the experiences and methods of the early adopters. Several
studies, e.g. (De Mauro et al., 2015; Wamba et al., 2015) recognize the need for guidelines and
a conceptual framework for big data. One way towards this goal is to examine the experiences
of real big data projects. In this article, we use computerized text analysis methods to analyze
a number of big data case studies documented in academic publications.
The key contribution of this article is that we synthesize the findings (benefits and
challenges) of our case study analysis to a set of generic themes and guidelines. This
contributes to the research on big data by conceptualizing existing practices and pointing out
several new research avenues. In addition, this work bridges practice and theory, acts as a
reference collection to currently known, peer-reviewed big data case studies, and benefits
practitioners by providing guidelines and experiences from the early adopters of big data.
2. Big Data Case Studies
This section describes the research process we used to identify big data case studies. We
used the systematic mapping study approach presented by Kitchenham (2007). Our goal was
to identify well-documented big data case studies in the academic literature. Well-documented
in this context means a peer-reviewed, high quality source. In order to cover the area broadly,
we performed a systematic mapping study. According to Kitchenham (2007), mapping studies
are designed to give a broad overview of a research area. Mapping studies have typically have
broad research questions. Our target (research question) was simple: to locate as many big data
case studies documented in peer-reviewed sources as possible, and to capture the common
concepts and lessons learned in these use cases. Figure 1 gives an overview of the search
process.
2.1 Search Strategy
Big data is a multi-disciplinary phenomenon. Unlike some other subject areas, big data
related articles cannot be found only in certain highly focused forums. Although there are some
new journals that focus on big data, various publications in many research domains discuss the
topic. Big data is an emerging, multi-disciplinary research area.
Initial search. First, in order to identify a representative set of well-documented studies,
we searched for cases in literature databases at the end of August 2015. We studied four
literature databases: Scopus, ProQuest, Web of Science, and EBSCO, using a rather broad
search terms (“’big data’ and ‘case study’”), and limiting our search to peer-reviewed papers.
The reason for this was to avoid “commercially-oriented” cases, i.e. we wanted to stick to
papers that had gone through scientific evaluation. In addition, we filtered the results to contain
only papers written in the English language. These searches revealed 281 papers.
3 (16)
Figure 1. Search process for big data case studies.
Exclusion criteria. First, we removed 115 duplicate papers from the initial result set. Then
we reviewed and categorized the remaining 166 articles. By reading the abstracts, and
whenever necessary the introduction and conclusions, we categorized the studies to be either
case-focused or non-case-focused. This led to the exclusion of 120 studies, which focused on
developing e.g. new algorithms, methods or frameworks. These papers verified or clarified
their contributions typically with an experimental prototype, proof-of-concept, or something
similar. Altogether, their focus was on developing something, not describing a case study. We
rejected additional 13 papers for various reasons: the paper contained a hypothetical case (3
studies), we could not access or find the paper (9), or the paper was in Spanish (1).
As a result, the search process revealed 33 peer-reviewed big data case study papers
containing in total 49 case studies due to three multi-case studies (Bärenfänger et al., 2014;
Kowalczyk and Buxmann, 2014; Wehn and Evers, 2015). Appendix 1 lists the papers and
provides a short contextual description of each paper. Next, we analyzed the articles describing
big data cases. For the analysis, we used a quantitative natural language processing software
to identify common concepts and themes. Finally, we analyzed the results of the text-mining
phase and formulated a set of guidelines.
2.2 Characteristics of the Case Studies
The found cases represented different application domains, from education to business, and
from healthcare to entertainment. This indicates that big data affects every aspect of life. Table
1 lists the number of cases categorized by the ISIC classification of the UN (UnitedNations,
2008). ISIC has 21 categories; we identified at least one big data case in 15 (71%) of these
categories. Transportation, especially intelligent transport systems -related studies, and various
healthcare studies represented the highest numbers of cases (8 and 6, respectively). Several
industries were also well represented with four or five cases each: manufacturing, retail,
finance, and information -related cases.
Table 1. Big data case studies by application area.
Application area (categories adopted from (UnitedNations, 2008))
Number of cases
A-Agriculture, forestry and fishing
1
B-Mining and quarrying
-
C-Manufacturing
5
D-Electricity, gas, steam and air conditioning supply
2
E-Water supply; sewerage, waste management and remediation activities
-
F-Construction
3
G-Wholesale and retail trade; repair of motor vehicles and motorcycles
5
H-Transportation and storage
8
4 (16)
I-Accommodation and food service activities
2
J-Information and communication
4
K-Financial and insurance activities
4
L-Real estate activities
-
M-Professional, scientific and technical activities
2
N-Administrative and support service activities
1
O-Public administration and defense; compulsory social security
1
P-Education
3
Q-Human health and social work activities
6
R-Arts, entertainment and recreation
2
S-Other service activities
-
T-Activities of households as employers; undifferentiated goods- and...
-
U-Activities of extraterritorial organizations and bodies
-
All the papers were recent, which is not surprising, since most organizations are still taking
their first steps with big data. Figure 2 presents the number of the case study articles per
publishing year of the paper (not the cases). Note that we did our searches at the end of August
2015, which explains the relatively low number of studies published in 2015.
Figure 2. Number of big data case study articles by the publishing year.
As with the application area, also the geographical distribution of the cases was wide,
representing five continents (figure 3). Companies based in North America and Europe
represented a majority of the cases with 29 instances. Beyond that, there were cases from Asia,
Australia and Africa. One of the studies, a multi-case study of 12 cases shown as “n/a” in figure
3, did not report the origin of the cases.
Figure 3. Geographical distribution of the big data cases.
Appendix 1 lists the case study papers. A brief description of the case context with industry
categorization provides basic information of the cases.
5 (16)
3. Content Analysis of the Case Study Papers
Content analysis is an established methodology for investigating textual data (see e.g.
(Berelson, 1952; Holsti, 1969; Krippendorf, 1989). Weber (1990) defines content analysis as
a repeatable, systematic procedure that reduces the many words of a text to much fewer content
categories. Novel applications of computerized content analysis have received the attention of
scholars recently, e.g. (Hu et al., 2014; Lewis et al., 2013; Yu et al., 2014), as researchers wish
to utilize new big data sources. In our case, manual coding of the texts of the 33 articles would
have been a time-consuming job, and therefore we considered computerized content analysis
to be a proper method for revealing common big data concepts and lessons learnt in the articles.
We had no pre-defined categories or themes. By using the data-driven approach, we just
drew the patterns from the articles with the analysis software. “Let the data speak”, as Mayer-
Schönberger and Cukier (2013) put it. As a tool we chose an open source software, KH Coder
2
.
It supports several text analysis methods described in content analysis studies, and more than
900 research projects have used the software.
Stemler (2001) suggests word counting and key words in context (KWIC) -analysis as a
starting point of a content analysis. KH Coder can count words, and more: it uses Stanford
POS Tagger (Toutanova et al., 2003) for tagging and lemmatization of words, i.e. it recognizes
parts of speech (such as nouns, verbs, and adjectives) and converts words to their base format.
This, combined to word frequency counting functionality, provided us a good basis for the
analysis. We used word frequencies and the KWIC analysis to create a so-called “stop-word”
list. Stop words are common words that exist in almost every sentence. Stop words are not
included in further analyses, as they do not add information; on the contrary, they make the
results more difficult to perceive. E.g. Wilbur and Sirotkin (1992), Yang and Pedersen (1997)
and Yang and Wilbur (1996) discuss automatic identification of stop words. We had to use the
manual method, as the KH Coder does not support automation. However, the KWIC analysis
tools of the KH Coder proved to be an efficient means to ensure whether the word was relevant
or not in the context that we were interested in.
We visualized the results with KH Coder software using co-occurrence maps.
3
Co-
occurrence maps build on the idea that words are related to the concepts they are connected to
(Ryan and Bernard, 2003). Osgood (1959) was among the first scholars to use co-occurrence
matrices to reveal connected concepts in textual data.
Figure 4 shows the co-occurrence map that resulted from the analysis of the 33 big data
case study articles (representing 49 cases) after several analysis iterations. The map revealed
five main themes and two sub-themes. The different colors distinguish the themes. We labelled
the themes based on the following: First, based on the virtual value creation process (Rayport
and Sviokla, 1995), we distinguished between data and data usage, as suggested by Ylijoki and
Porras (2016). Three of the main themes are business or organization -related, representing the
usage or utilization of data. Two of them are ICT- and data-related, technical themes. Then we
2
KH Coder is a free software for quantitative content analysis or text mining - http://khc.sourceforge.net/en/
3
A few notes that clarify the interpretation of the map: When plotting a map, the KH Coder uses the method
explained in (Fruchterman and Reingold, 1991). This algorithm may plot nodes side by side, but unlike e.g.
multi-dimensional maps, this does not necessarily indicate co-occurrence. Instead, edges (lines) indicate co-
occurrence: if a line connects the nodes (words), co-occurrence exists. For example, in figure 4, the terms
‘customer’ and ‘organization’ are close to each other, but there is no co-occurrence between them, since there is
no line between the words. Accordingly, a strong co-occurrence between the terms ’value’ and ‘generate’
exists, as there is a thick line between them. The thicker the line, the stronger the co-occurrence is. The dotted
lines show co-occurrence between terms that belong to different communities (i.e. themes). The size of the plot
indicates the frequency of the term, 'data' being obviously the term used most frequently in the articles. The
color-coding indicates the communities (sub-graphs) that are relatively close to each other. The KH Coder
offers several methods for indicating patterns. We used the modularity method defined in (Clauset et al., 2004).
This method builds on the principle that there are many edges within the communities and only a few between
them.
6 (16)
decided the label for each theme based on KWIC-analysis and manual inspection of the
articles. The co-occurrence of words within a theme is presented with a solid line between the
words.
Figure 4. Co-occurrence map of the terms in big data case study articles.
The three business (or data usage) -related themes are:
Decision-making (red color in the map). Several studies discussed enhancing the
decision-making processes, enabling data-driven decision-making, or providing
actionable insights to managers (Bärenfänger et al., 2014; Dutta and Bose, 2015;
Krumeich et al., 2014). Several studies (Cai et al., 2014; Tao et al., 2014; Kalakou et
al., 2015) also investigated transportation or passenger patterns, providing insights into
planning and decision-making. Embedding analytics and insights into processes and
decision-making routines is important (Bekmamedova and Shanks, 2014). However,
according to the case studies, there are challenges to overcome in this area, such as lack
of data-driven organizational culture (Shen and Varvel, 2013; Dutta and Bose, 2015),
missing analytics strategy, and lack of leadership (Phillips-Wren and Hoskisson, 2015).
Innovation (blue). Big data was seen as an enabler for data-driven innovation and
faster innovation cycles (Amatriain, 2013; Jetzek et al., 2014). In addition, (Martinez
and Walton, 2014) reported successful and cost-efficient usage of crowd-sourced big
data analytics, and (Ciulla et al., 2012) used social media data to predict the winner of
a song contest.
Business value (light yellow). According to the studies, big data is a vehicle to create
new value. The studies recognized positive results and opportunities, such as a business
7 (16)
model that was based on big data (Amatriain, 2013), energy and cost savings (Dobson
et al., 2014; Jetzek et al., 2014; Mathew et al., 2015), business transformation
(Prescott, 2014), increased revenue and customer satisfaction (Dutta and Bose, 2015),
better transparency over operations (Bärenfänger et al., 2014), generating value by
secondary use of data (Bettencourt-Silva et al., 2015), and deeper understanding of real
events (Crampton et al., 2013; Hu et al., 2014). The other side of this coin is that there
are challenges related to the technical themes.
The two ICT-related themes cover data and analytics, new data sources, and data
management aspects.
Data management (cyan) through the whole lifecycle of data, from the sources to the
analytics, is a central aspect. In general, the volume, variety, and velocity of big data
can be challenging for data management and technology (Laney, 2001). Companies are
experimenting with new technologies (Bärenfänger et al., 2014). Some studies
mentioned that managing the volumes of data is a key challenge (Krumeich et al., 2014;
Dutta and Bose, 2015). Moreover, the case studies pointed out additional aspects that
need to be addressed, such as data inconsistencies and poor data quality (O’Leary,
2013a; Halamka, 2014; Mathew et al., 2015). Several studies also reported concerns
for potential security and/or privacy issues (Halamka, 2014; Martinez and Walton,
2014; Stephansen and Couldry, 2014; Bettencourt-Silva et al., 2015). Applying proper
analytics to the vast amounts of data is the key in gaining value and insights. New data
types, such as social media posts or text documents, require new kinds of analytics.
This is a multi-faceted issue: in addition to new technology, organizations need new
talent, both business-oriented and technology-skilled (Shen and Varvel, 2013; Phillips-
Wren and Hoskisson, 2015; Prinsloo et al., 2015).
New data sources (purple). In several cases organizations utilized data from outside
their own organization, such as Facebook and Twitter data (He et al., 2013), blog texts
and user reviews (Marine-Roig and Clavé, 2015), or data collected from mobile apps
(O’Leary, 2013a; Papenfuss et al., 2015). They had been able to extract value from
these external sources. The data are freely available, but requires quite a lot of
processing, as described e.g. in (Marine-Roig and Clavé, 2015).
In addition to the five main themes, the map in figure 4 shows a few sub-themes. The
Pattern-identify theme is related to data. The KWIC-analysis showed that these keywords were
mostly used when the articles discussed revealing patterns from data. The User-tweet theme
rose from articles in which Twitter-analyses were discussed. The keyword 'new' was used in
various contexts, but mostly in conjunction with data. Accordingly, the keyword 'customer'
was mostly used in contexts that discussed a company’s customer insights.
The map also shows several dotted lines between the words that belong to different themes.
This indicates that the themes are inter-related. For example, there are several relations
between the nodes that belong to decision-making, business value and data management
themes. Concrete indications of these relations are the challenges regarding decision-making
and data management. The linkages reflect the disruptive impact of big data and the inevitable
business transformation process. The case study articles provide minimal information on how
to solve the challenges, which opens new research avenues like those listed in the conclusions
section. Data management theme inter-relates also with new data sources theme. This is
intuitively obvious. However, as many organizations lack the required analytical and technical
capabilities, they will turn to external vendors, and new data or analytics related services will
emerge.
8 (16)
4. Discussion and Lessons Learned
The themes we discovered pointed out three essential business aspects – decision-making,
innovation, and business value – related to big data. Regarding these themes, many of the cases
reported positive results. However, to meet the big data value proposal discussed in the
Introduction section, business transformation and new business models are required. We could
identify in the articles one case where the business model was based on big data (Amatriain,
2013): Netflix runs their business based on the data they collect, and boosts their sales by
making customer-specific, data-driven recommendations. One case study (Prescott, 2014)
reported a business transformation process leading to re-gaining competitive advantage. There
was also one case where a company-wide data-driven approach was taken (Dutta and Bose,
2015). In this case, a large steel manufacturer reshaped their processes and functions to take
advantage of data, which resulted in significant business benefits. On the other hand, they also
faced challenges, such as organizational resistance towards the change. The rest of the cases
were more function-specific, limited-scope initiatives that brought benefits to certain
operations, e.g. marketing, or social media -related experiments. Several cases, e.g. (Crampton
et al., 2013; Cheng and Chen, 2014; Hu et al., 2014) had analyzed social media data in order
to identify signs or clues of e.g. raising trends or other emerging actions. For a discussion of
Internet of Signs, see (O’Leary, 2013b). One aspect in data-driven innovation is secondary use
of data, which means that the data are used to another purpose than it was originally collected
for. Some of the sources, e.g. (Mayer-Schönberger and Cukier, 2013) claimed that the
secondary usage of data has huge potential. We identified one case (Bettencourt-Silva et al.,
2015), where this kind of data usage was clearly recognized and utilized. Of course, these
findings must be compared against the fact that most of the organizations were taking their
first steps on the big data path.
Technology and software vendors typically emphasize the business aspects, and especially
their positive effects. As our results point to the same direction, our study confirms the hype
partly. What the hype typically leaves out is that changing the organization to a more data-
driven one will have effects on the organizational culture and decision-making processes, as
the challenges related to the decision-making theme indicate. Moreover, several studies
reported technical challenges, especially with the data volumes.
Table 2 synthesizes the findings of our analysis. The examples column includes examples
of the articles related to the theme. The case studies showed that the value proposal of big data
is significant. However, realizing the value is much more a business transformation initiative
than a technical issue. Organizations need to consider these aspects carefully in their big data
experiments.
Table 2. Guidelines for big data utilization.
Theme
Guidelines
Examples
Decision-
making
Embed analytics into decision-making
processes.
Be prepared for organizational side
effects.
(Bekmamedova and Shanks,
2014), (Cai et al., 2014), (Dutta
and Bose, 2015), (Phillips-Wren
and Hoskisson, 2015)
Innovation
Trust the data.
Search for new methods.
(Amatriain, 2013), (Jetzek et al.,
2014), (Martinez and Walton,
2014), (Ciulla et al., 2012)
Business
value
Look for value in various directions;
experiment with the data.
(Amatriain, 2013), (Bettencourt-
Silva et al., 2015), (Dutta and
9 (16)
Enable business transformation with
the data.
Consider secondary usage of the data.
Bose, 2015), (O’Leary, 2013a),
(Prescott, 2014)
Data
management
Expect to face technical and data-
related challenges.
Plan for security.
(Dutta and Bose, 2015),
(Halamka, 2014), (Krumeich et
al., 2014), (Prinsloo et al.,
2015), (Shen and Varvel, 2013)
New data
sources
Experiment with new data types.
Consider potential privacy issues.
(He et al., 2013), (Marine-Roig
and Clavé, 2015), (Yu et al.,
2014)
Data and analytics should be embedded into the decision-making processes. Taking
advantage of analytic software suggestions and decision support information should be a habit
in a data-driven organization. This is possible only if the information is easily available in the
normal decision-making context. A recent study suggests that tight integration to enterprise
systems is a success factor to business intelligence solutions (Isik et al., 2011). However, this
can be a difficult task. For example, middle management and specialists make important
operative decisions. Although the cases did not discuss this matter, it is obvious that
embedding analytics into their working context and to legacy systems can be a complex and
expensive task. It would require significant changes to legacy systems, often combining new
and old technologies. Another aspect to consider is that the organizational side effects of the
data-driven approach can be significant. Several of the case studies reported challenges in this
area. In order to gain benefits, a data-driven organizational attitude is required, but the
organizational culture often hinders the change. In addition, utilizing data may lead to changes
in the decision-making processes. Managers need to not only understand but also support these
changes. Managing the change and the organizational side effects requires training and new
managerial skills.
The data-driven innovation method requires rapid testing of many new hypotheses and
ideas, gathering data from the tests and – most importantly – relying on the data that results
from the tests. This kind of process is described in (Amatriain, 2013). Netflix runs several tests
simultaneously in order to improve their services. Although in this case the services are digital,
the principle is general. Instead of concentrating on finalizing one solution at a time, a better
approach might be to test several primitive prototypes with the customers at the same time.
The feedback would help to improve the solution, to ensure that the solution really is
something that the customers need, and to speed up the innovation pace (Furr and Dyer, 2014).
However, relying on data and an experimental, more customer-centric innovation method
requires the organizational culture to allow mistakes and uncertainty. Many ideas simply do
not make it, and the more disruptive the idea, the more difficult it is to calculate the business
case.
The business value of big data value potential is case-dependent. According to our analysis
of the case studies, big data can drive business value and innovation. The cases reported
various opportunities in different areas. However, the opportunities are case-dependent, so
each organization must do their thinking in order to find out how to add value with data. What
is the business problem that we are trying to solve with big data? One important aspect to
consider is the secondary usage of data. As organizations generate and harvest more and more
data, opportunities will open to utilize the data in new, unexpected ways that can generate
value. For example, a factory that must collect real-time emission data for regulatory purposes
might be able to use the same data for another purpose, such as process monitoring.
10 (16)
The data management challenge of big data is real. Several of the studies reported
significant issues with data volumes. New technologies are rapidly emerging, and
organizations should be able to integrate these into their current infrastructure. This requires
architectural and technical talent, money, and company policies that allow new vendors and
technologies to enter to the playground. Security issues are obvious: where there is value, there
is a potential fraud. Data protection must be secured from the source to the presentation.
Security must be planned and built into the systems. Many of the case studies recognized
potential problems in these areas. The case studies also recognized challenges in data quality
and the shortage of analytic capabilities. These are partly technical issues, but they also require
business talent.
New data sources can provide value. Several of the case studies mined out value from
tweets or other textual data, as did we. Our own experiments with the computerized content
analysis suggested that appropriate software tools are efficient and cost-effective (compared
to manual coding). This makes content analysis a viable option also for practitioners. The main
caveat here is that text analysis requires knowledge in the theory and methods of content
analysis. Another consideration are the tools. According to (Isik et al., 2011), users are
dissatisfied with external data capabilities of current tools. However, integrating new, external
data sources would also improve user satisfaction. From the privacy point of view, combining
analytics and data from several sources can lead to unpredicted privacy issues. Companies
must consider the public opinion as well as the governing policies and legislation.
5. Conclusions
Several studies, e.g. (Manyika et al., 2011; Mayer-Schönberger and Cukier, 2013;
Davenport, 2014) have made claims that big data causes pervasive changes, which will affect
almost every sector of life. In this study, we analyzed 33 peer-reviewed papers describing 49
big data use cases. The cases confirmed the claims, at least partly. Clearly, big data applications
are emerging in various areas of life. The studies recognized positive results and opportunities,
such as new business models, energy and cost savings, cost-efficient open innovation, business
transformation, or deeper understanding of real events for decision support. Previous research,
like McAfee and Brynjofsson (2012), have shown that data driven decisions add value to the
business. Our research used a different methodology and a different research set, but the results
point to the same direction, supporting the results of the previous research.
However, several studies also reported of challenges like data inconsistencies and poor data
quality, security and/or privacy issues, missing analytics strategy, lack of leadership, lack of
data-driven organizational culture, and the need of new analytics and technology skills. These
challenges reflect the disruptive nature of big data. They are indications of major shifts
required; changes that affect not only technical platforms and skills, but they also – and more
importantly – influence the organizational culture, decision-making processes and
management functions. Previous studies discuss many of these challenges in general level.
Based on current big data implementations as described on peer-reviewed literature, our study
adds insights at more concrete level, providing practitioners best practices and guidance to
avoid common pitfalls.
We used computerized content analysis to extract knowledge from the raw text of the case
study papers. Using the computerized approach with open-source tools enables organizations
to experiment with text analysis. The results of the computer analysis must be processed further
and proved to be useful. We interpreted the results of a co-occurrence map to five named
themes and verified the results against case study papers. These insights enabled us to
conceptualize the findings to a set of guidelines (see Table 2) that point out several essential
aspects that organizations must consider in their big data experiments. These guidelines
11 (16)
emphasize that dealing with big data is a complicated task, which requires addressing
technical, business-related and organizational issues.
In this research we created a set of guidelines stating what organizations should consider
when dealing with big data. Another viewpoint is how to tackle the topics. This is an important
question especially for practitioners. However, only a few articles discussed the case studies
on a detailed enough level to answer the how question. This opens new research avenues. We
point out some of these avenues below.
For researchers focusing on big data topics in the business context, this study offers a
collection of big data case studies to start with, and several possibilities for further research.
In addition to several technical questions, there are many open questions related to business
transformational effects of big data, including the following.
Understanding business transformation processes behind digitalization and big data.
How does datafication drive the change in different industries? How can an
organization adapt to the changes in industry structures and ecosystems? What is
required to manage the change effectively?
What are the effects of big data on the decision-making processes of the organization?
What organizational effects does this have? How should an organization integrate big
data analytics effectively to the existing business processes and workflows?
How does big data enable innovation? What are the driving forces behind the new,
data-based innovation processes? How should an organization arrange its innovation
method to be effective in the big data era?
What methods and processes are efficient when organizations start to explore big data?
How do the existing infrastructure and company policies match with big data
experimenting? How could companies evaluate various options quickly in order to
decide which of them are promising, and what kind of risks they contain?
6. References
Amatriain, X., 2013. Beyond data: from user information to business value through
personalized recommendations and consumer science. Proceedings of the 22nd ACM
international conference on information & knowledge management 2201–2208.
Bärenfänger, R., Otto, B., Österle, H., 2014. Business value of in-memory technology–
multiple-case study insights. Industrial Management & Data Systems 114, 1396–
1414.
Bekmamedova, N., Shanks, G., 2014. Social Media Analytics and Business Value: A
Theoretical Framework and Case Study. System Sciences (HICSS), 2014 47th Hawaii
International Conference on 3728–3737.
Berelson, B., 1952. Content analysis in communication research. US Free Press, New York.
Bettencourt-Silva, J.H., Clark, J., Cooper, C.S., Mills, R., Rayward-Smith, V.J., De La
Iglesia, B., 2015. Building Data-Driven Pathways From Routinely Collected Hospital
Data: A Case Study on Prostate Cancer. JMIR medical informatics 3, 1–21.
Cai, H., Jia, X., Chiu, A.S., Hu, X., Xu, M., 2014. Siting public electric vehicle charging
stations in Beijing using big-data informed travel patterns of the taxi fleet.
Transportation Research Part D: Transport and Environment 33, 39–46.
Cheng, Y.-C., Chen, P.-L., 2014. Global social media, local context: A case study of
Chinese-language tweets about the 2012 presidential election in Taiwan. Aslib
Journal of Information Management 66, 342–356.
Christensen, C., 2013. The innovator’s dilemma: when new technologies cause great firms to
fail. Harvard Business Review Press.
12 (16)
Ciulla, F., Mocanu, D., Baronchelli, A., Gonçalves, B., Perra, N., Vespignani, A., 2012.
Beating the news using social media: the case study of American Idol. EPJ Data
Science 1, 1–11.
Clauset, A., Newman, M.E., Moore, C., 2004. Finding community structure in very large
networks. Physical review E 70, 1–6.
Crampton, J.W., Graham, M., Poorthuis, A., Shelton, T., Stephens, M., Wilson, M.W., Zook,
M., 2013. Beyond the geotag: situating “big data”and leveraging the potential of the
geoweb. Cartography and geographic information science 40, 130–139.
Davenport, T., 2014. Big data at work: dispelling the myths, uncovering the opportunities.
Harvard Business Review Press.
De Mauro, A., Greco, M., Grimaldi, M., 2015. What is big data? A consensual definition and
a review of key research topics, in: AIPConferenceProceedings. pp. 97–104.
Dehning, B., Richardson, V.J., Zmud, R.W., 2003. The value relevance of announcements of
transformational information technology investments. Mis Quarterly 27, 637–656.
Dobson, G., Tilson, D., Tilson, V., Haas, C.E., 2014. Quantitative case study: Use of
pharmacy patient information systems to improve operational efficiency. System
Sciences (HICSS), 2014 47th Hawaii International Conference on 4220–4228.
Dutta, D., Bose, I., 2015. Managing a Big Data project: The case of Ramco Cements
Limited. International Journal of Production Economics 165, 293–306.
E. Prescott, M., 2014. Big data and competitive advantage at Nielsen. Management Decision
52, 573–601.
Fang, S., Da Xu, L., Zhu, Y., Ahati, J., Pei, H., Yan, J., Liu, Z., 2014. An integrated system
for regional environmental monitoring and management based on internet of things.
Industrial Informatics, IEEE Transactions on 10, 1596–1605.
Fruchterman, T.M., Reingold, E.M., 1991. Graph drawing by force-directed placement.
Softw., Pract. Exper. 21, 1129–1164.
Furr, N., Dyer, J., 2014. The Innovator’s Method. Harvard Business Review Press.
Gantz, J., Reinsel, D., 2011. Extracting value from chaos (No. 1142), IDC iview. IDC.
Halamka, J.D., 2014. Early Experiences with big data at an academic medical center. Health
affairs 33, 1132–1138.
He, W., Zha, S., Li, L., 2013. Social media competitive analysis and text mining: A case
study in the pizza industry. International Journal of Information Management 33,
464–472.
Holsti, O.R., 1969. Content analysis for the social sciences and humanities. Addison-Wesley.
Hu, H., Ge, Y., Hou, D., 2014. Using web crawler technology for geo-events analysis: A
case study of the Huangyan Island incident. Sustainability 6, 1896–1912.
Isik, O., Jones, M.C., Sidorova, A., 2011. Business intelligence (BI) success and the role of
BI capabilities. Intelligent systems in accounting, finance and management 18, 161–
176.
Jetzek, T., Avital, M., Bjorn-Andersen, N., 2014. Data-driven innovation through open
government data. Journal of theoretical and applied electronic commerce research 9,
100–120.
Kalakou, S., Psaraki-Kalouptsidi, V., Moura, F., 2015. Future airport terminals: New
technologies promise capacity gains. Journal of Air Transport Management 42, 203–
212.
Kitchenham, B., 2007. Guidelines for performing systematic literature reviews in software
engineering (No. EBSE-2007-01), Technical report, Ver. 2.3 EBSE Technical Report.
EBSE. Keele University.
Kolowitz, B.J., Lauro, G.R., Venturella, J., Georgiev, V., Barone, M., Deible, C., Shrestha,
R., 2014. Clinical Social Networking—A New Revolution in Provider
Communication and Delivery of Clinical Information across Providers of Care?
Journal of digital imaging 27, 192–199.
Kowalczyk, M., Buxmann, P., 2014. Big Data and Information Processing in Organizational
Decision Processes. Business & Information Systems Engineering 6, 267–278.
13 (16)
Krippendorf, K., 1989. Content analysis. International ensyclopedia of communication 1,
403–407.
Krumeich, J., Jacobi, S., Werth, D., Loos, P., 2014. Towards planning and control of
business processes based on event-based predictions, in:
BusinessInformationSystems. Springer, pp. 38–49.
Laney, D., 2001. 3D data management: Controlling data volume, velocity and variety. META
Group Research Note 6, 70.
Lewis, S.C., Zamith, R., Hermida, A., 2013. Content analysis in an era of big data: A hybrid
approach to computational and manual methods. Journal of Broadcasting &
Electronic Media 57, 34–52.
Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C., Byers, A.H., 2011.
Big data: The next frontier for innovation, competition, and productivity. McKinsey
Global Institute.
Marine-Roig, E., Clavé, S.A., 2015. Tourism analytics with massive user-generated content:
A case study of Barcelona. Journal of Destination Marketing & Management.
doi:http://dx.doi.org/10.1016/j.jdmm.2015.06.004i
Martinez, M.G., Walton, B., 2014. The wisdom of crowds: The potential of online
communities as a tool for data analysis. Technovation 34, 203–214.
Mathew, P.A., Dunn, L.N., Sohn, M.D., Mercado, A., Custudio, C., Walter, T., 2015. Big-
data for building energy performance: Lessons from assembling a very large national
database of building energy use. Applied Energy 140, 85–93.
Mayer-Schönberger, V., Cukier, K., 2013. Big data: A revolution that will transform how we
live, work, and think. Houghton Mifflin Harcourt.
McAfee, A., Brynjolfsson, E., 2012. Big data: The Management Revolution. Harvard
Business Review 90, 61–67.
O’Leary, D.E., 2013a. Exploiting big data from mobile device sensor-based apps: Challenges
and benefits. MIS Quarterly Executive 12, 179–187.
O’Leary, D.E., 2013b. Big Data, the Internet of Things and Internet of Signs. Intelligent
Systems in Accounting, Finance and Management 20, 53–65.
Osgood, C.E., 1959. The representational model and relevant research methods. Trends in
content analysis 33–88.
Papenfuss, J.T., Phelps, N., Fulton, D., Venturelli, P.A., 2015. Smartphones Reveal Angler
Behavior: A Case Study of a Popular Mobile Fishing Application in Alberta, Canada.
Fisheries 40, 318–327.
Phillips-Wren, G., Hoskisson, A., 2015. An analytical journey towards big data. Journal of
Decision Systems 24, 87–102.
Porter, M.E., Millar, V.E., 1985. How information gives you competitive advantage.
Prescott, M., 2014. Big data and competitive advantage at Nielsen. Management Decision
52, 573–601.
Prinsloo, P., Archer, E., Barnes, G., Chetty, Y., Van Zyl, D., 2015. Big (ger) data as better
data in open distance learning. The International Review of Research in Open and
Distributed Learning 16.
Ryan, G.W., Bernard, H.R., 2003. Techniques to identify themes. Field methods 15, 85–109.
Sainio, L.-M., 2005. The Effects of Potentially Disruptive Technology on Business Model -
A Case Study of New Technologies in ICT Industry. Lappeenranta University of
Technology.
Shen, Y., Varvel, V.E., 2013. Developing data management services at the Johns Hopkins
University. The Journal of Academic Librarianship 39, 552–557.
Stemler, S., 2001. An overview of content analysis. Practical assessment, research &
evaluation 7, 137–146.
Stephansen, H.C., Couldry, N., 2014. Understanding micro-processes of community building
and mutual learning on Twitter: a “small data”approach. Information, Communication
& Society 17, 1212–1227.
14 (16)
Tao, S., Corcoran, J., Mateo-Babiano, I., Rohde, D., 2014. Exploring Bus Rapid Transit
passenger travel behaviour using big data. Applied Geography 53, 90–104.
Toutanova, K., Klein, D., Manning, C.D., Singer, Y., 2003. Feature-rich part-of-speech
tagging with a cyclic dependency network. Proceedings of the 2003 Conference of
the North American Chapter of the Association for Computational Linguistics on
Human Language Technology-Volume 1 173–180.
UnitedNations, 2008. International Standard Industrial Classification of All Economic
Activities. United Nations.
Venkatraman, N., 1994. IT-enabled business transformation: from automation to business
scope redefinition. Sloan management review 35, 73–73.
Wamba, S.F., Akter, S., Edwards, A., Chopin, G., Gnanzou, D., 2015. How “big data”can
make big impact: Findings from a systematic review and a longitudinal case study.
International Journal of Production Economics 165, 234–246.
Weber, R.P., 1990. Basic content analysis. Sage.
Wehn, U., Evers, J., 2015. The social innovation potential of ICT-enabled citizen
observatories to increase eParticipation in local flood risk management. Technology
in Society 42, 187–198.
Wilbur, W.J., Sirotkin, K., 1992. The automatic identification of stop words. Journal of
information science 18, 45–55.
Yang, Y., Pedersen, J.O., 1997. A comparative study on feature selection in text
categorization, in: ICML. pp. 412–420.
Yang, Y., Wilbur, J., 1996. Using corpus statistics to remove redundant words in text
categorization. JASIS 47, 357–369.
Ylijoki, O., Porras, J., 2016. Perspectives to Definition of Big Data: A Mapping Study and
Discussion. Journal of Innovation Management (in press).
Yu, K., Zhang, J., Chen, M., Xu, X., Suzuki, A., Ilic, K., Tong, W., 2014. Mining hidden
knowledge for drug safety assessment: topic modeling of LiverTox as a case study.
BMC bioinformatics 15, S6.
15 (16)
Appendix 1 – Big Data Case
Study Articles
Table 1 is a summary of the case study articles we analyzed. The table contains 33 different
studies and 49 big data cases (due to 3 multi-case studies) representing five continents. We
identified academic, peer-reviewed articles in major literature databases (ProQuest, SCOPUS,
Web-of-Science, and EBSCO) covering business and technical topics at the end of August
2015.
The context column describes the focus area of the study. Application area is the
categorization of the case(s) that the article reports, according to ISIC classification
(UnitedNations, 2008). Country is the origin of the organization subject to the study.
Appendix 1 – Table 1. Big data case study articles.
Paper
Context
Application area
(ISIC)
Country
(Amatriain, 2013)
Netflix recommender system.
J-Information and
communication
USA
(Bekmamedova
and Shanks, 2014)
Marketing campaign using social media.
K-Financial and
insurance activities
Australia
(Bettencourt-Silva
et al., 2015)
Secondary usage of routinely collected patient
data.
Q-Human health and
social work activities
UK
(Bärenfänger et
al., 2014)
In-memory computing business value assessed
in 5 large European companies from different
industries.
C-Manufacturing (2)
G-Wholesale and retail
D-Electricity
H-Transportation
n/a
(Europe)
(Cai et al., 2014)
Taxi trajectory data used to reveal travel patterns
in order to help the planning of public charging
infrastructure.
H-Transportation
China
(Cheng and Chen,
2014)
Analysis of Twitter communities during the
presidential election in Taiwan in 2012.
O-Public
administration
Taiwan
(Ciulla et al.,
2012)
Predicting the American Idol competition results
by using Twitter analysis.
R-Arts, entertainment
and recreation
USA
(Crampton et al.,
2013)
Social and spatial analysis of geotagged tweets
following the 2012 NCAA championships.
R-Arts, entertainment
and recreation
USA
(Dobson et al.,
2014)
Cost reductions in a hospital by process
analytics.
Q-Human health and
social work activities
USA
(Dutta and Bose,
2015)
Big data initiative in a manufacturing company.
C-Manufacturing
India
(Fang et al., 2014)
An integrated system for monitoring regional
environmental data (collecting, storing and
analyzing temperature-related data).
M-Professional,
scientific and technical
activities
China
(Martinez and
Walton, 2014)
By adopting a crowdsourcing approach to data
analysis (using Kaggle), Dunnhumby were able
to extract information from their own data that
was previously unavailable to them.
G-Wholesale and retail
UK
(Halamka, 2014)
Analysis and experiences of new big data
possibilities and challenges in a hospital.
Q-Human health and
social work activities
USA
(He et al., 2013)
Social media marketing in the pizza industry.
G-Wholesale and retail
USA
(Hu et al., 2014)
The Huangyan Island incident was studied by
using a web crawler technology and text
analysis.
M-Professional,
scientific and technical
activities
(Huangyan
Island,
South China
Sea)
(Jetzek et al.,
2014)
Case Opower: generating value from open data.
Saving energy by offering benchmark
information to consumers.
D-Electricity, gas,
steam and air
conditioning supply
USA
(Kalakou et al.,
2015)
Simulation for planning airport terminals and
reducing passenger check-in and security
H-Transportation and
storage
Portugal
16 (16)
Paper
Context
Application area
(ISIC)
Country
checkpoint times, using Lisbon airport as the
case.
(Kolowitz et al.,
2014)
Using social technologies to construct dynamic
provider networks, simplify communication, and
facilitate clinical workflow operations.
Q-Human health and
social work activities
USA
(Kowalczyk and
Buxmann, 2014)
Multi-case study, 12 big companies from
various industries.
J-Information and
communication (2)
K-Financial and
insurance activities (3)
G-Wholesale and retail
(2)
I-Accommodation and
food service activities
H-Transportation and
storage (2)
Q-Human health and
social work activities
C-Manufacturing
n/a
(Krumeich et al.,
2014)
Big data experiments and challenges of a steel
factory.
C-Manufacturing
Germany
(Lewis et al.,
2013)
A case of news sourcing on Twitter combining
text mining and manual methods.
J-Information and
communication
USA
(Marine-Roig and
Clavé, 2015)
Tourism and city strategy planning and
marketing in the Barcelona region by using big
data analytics.
N-Administrative and
support service
activities
Spain
(Mathew et al.,
2015)
Case study of the largest database of building
energy data in US; aiming at enabling energy
savings.
F-Construction
USA
(O’Leary, 2013a)
A mobile device application collecting data that
the city of Boston uses to facilitate road
infrastructure management.
H-Transportation and
storage
USA
(Papenfuss et al.,
2015)
Analyzing behavioral patterns in fishing by
using mobile app -generated data.
A-Agriculture, forestry
and fishing
Canada
(Phillips-Wren and
Hoskisson, 2015)
Case Choice-hotels customer analytics (CRM,
Twitter).
I-Accommodation and
food service activities
USA
(E. Prescott, 2014)
Nielsen re-gaining their competitive advantage
by using data and analytics.
H-Transportation and
storage
USA
(Prinsloo et al.,
2015)
Unifying and analyzing data (360 000 students,
courses, programs etc.) at the University of
South Africa (Unisa).
P-Education
South
Africa
(Shen and Varvel,
2013)
New data management services platform
implementation at Johns Hopkins University.
Aims to increase data and knowledge sharing.
P-Education
USA
(Stephansen and
Couldry, 2014)
A case study where a departmental Twitter
account was used to create a community of
practice (students and teachers) and to enable
mutual learning beyond the classroom.
P-Education
UK
(Tao et al., 2014)
Big data visualization case in bus-rapid-transit in
order to understand passenger travel dynamics
and plan capacity.
H-Transportation and
storage
Australia
(Wehn and Evers,
2015)
Planning and managing flooding situations, 2
cases
F-Construction
UK
Netherlands
(Yu et al., 2014)
Text mining (topic modeling) applied to text
documents in order to improve drug safety by
finding drugs susceptible to acute liver failures.
Q-Human health and
social work activities
USA
