Wright State University

CORE Scholar

Information Systems and Supply Chain

ISSCM Faculty Publications Management

2013

Research and Pedagogy in Business Analytics: Opportunities and

Illustrative Examples
Ramesh Sharda

Daniel Asamoah
Wright State University - Main Campus, [email protected]

Natraj Ponna

Follow this and additional works at: https://corescholar.libraries.wright.edu/infosys_scm

Part of the Management Information Systems Commons, Operations and Supply Chain Management
Commons, and the Social Media Commons

Repository Citation
Sharda, R., Asamoah, D., & Ponna, N. (2013). Research and Pedagogy in Business Analytics: Opportunities
and Illustrative Examples. Journal of Computing and Information Technology, 21 (3), 171-183.
https://corescholar.libraries.wright.edu/infosys_scm/4

This Article is brought to you for free and open access by the Information Systems and Supply Chain Management
at CORE Scholar. It has been accepted for inclusion in ISSCM Faculty Publications by an authorized administrator
of CORE Scholar. For more information, please contact [email protected].
Journal of Computing and Information Technology - CIT 21, 2013, 3, 171–183 171
doi:10.2498 /cit.1002194

Research and Pedagogy in Business

Analytics: Opportunities and
Illustrative Examples

Ramesh Sharda, Daniel Adomako Asamoah and Natraj Ponna

Institute for Research in Information Systems, Spears School of Business, Oklahoma State University, USA

In recent times, business analytics and big data have a sense of what is happening with an organiza-
gained momentum both in industry practice and academic tion. Additional technologies have enabled us to
research. The objective of this paper is to provide both a
research and teaching introduction to business analytics create more customized reports that can be gen-
in the context of both current and prospective perspective erated on an ad hoc basis. The next extension
of the business analytics domain. It begins by providing of reporting takes us to online analytical pro-
a quick overview of the three types of analytics. To assist cessing (OLAP) type queries that allow users
the future analytics professionals, we identify, group
and discuss nine different participants of the analytics to dig deeper and determine specific sources of
industry into clusters. We then include a brief description concern or opportunities. Technologies avail-
of some current research projects under way in our team. able today can also automatically issue alerts
We also note some research opportunities in Big Data
analytics. The paper also concludes with a discussion of for decision makers when the performance is-
teaching opportunities in analytics. sues warrant such alerts. At a consumer level,
we see such alerts for weather or other issues.
Keywords: business analytics, Big Data, social media But similar alerts can also be generated in spe-
research, athletic injuries, healthcare cific settings, for example, when the sales fall
above or below a certain level within a certain
time period, or when the inventory for a specific
product is running low. All of these applications
1. Introduction are made possible through analysis and queries
on data being collected by an organization. The
next level of analysis entails statistical analysis
Although many authors and consultants have
defined it slightly differently, analytics can be to better understand patterns, which can be used
viewed as the process of developing recommen- to develop forecasts or models for predicting
how customers might respond to a specific mar-
dations for actions or actionable decisions based
keting campaign or ongoing service/product
upon insights generated from historical data.
offering. When an organization has a good
The Institute for Operations Research and Man-
view of what is happening and what is likely
agement Science (INFORMS) has created a ma-
to happen, it can also employ other techniques
jor initiative to organize and promote analyt-
to make the best decisions under the circum-
ics. According to INFORMS, analytics repre-
stances. These eight levels of analytics are de-
sents the combination of computer technology,
scribed in more detail in a white paper by SAS.
management science techniques, and statistics
(http://www.sas.com/news/sascom/analy-
to solve real problems. Of course, many other
tics levels.pdf) [13].
organizations have proposed their own interpre-
tations and motivation for analytics. For exam- Instead of eight levels, INFORMS defines three
ple, SAS Institute Inc. proposed eight levels of categories of analytics, based on the idea of
analytics that begin with standardized reports looking at all the data to understand what is
from a computer system. These reports provide happening, what will happen, and how to make
172 Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples

the best of it. These three levels are identified customer is a credit risk, etc. A number of tech-
(http://www.informs.org/Community/Ana- niques are used in developing predictive analyti-
lytics) as Descriptive, Predictive, and Pre- cal applications including various classification
scriptive [5]. Figure 1 presents a graphical view algorithms. For example, we can use classifi-
of these three levels of analytics. The intercon- cation techniques such as decision tree models
nected circles show that there is some overlap and neural networks to predict how well a mo-
across these three types of analytics. A more tion picture would do at the box office. We can
detailed description of each of these categories also use clustering algorithms for segmenting
follows. customers into different clusters to be able to
target specific promotions to them. Finally, we
can use association mining techniques to esti-
mate relationships between different purchasing
behaviors. That is, if a customer buys one prod-
uct, we can predict other items the customer is
likely to purchase. Such analysis can assist a
retailer in recommending or promoting related
products. For example, any product search on
Amazon.com results in the retailer also suggest-
ing other similar products that a customer may
be interested in.
The third category of analytics is termed Pre-
scriptive analytics. The goal of prescriptive an-
alytics is to examine current trends and likely
Figure 1. Three types of analytics. forecasts and use that information to make deci-
sions. This group of techniques has historically
been studied under the umbrella of operations
Descriptive or reporting analytics refers to know- research or management sciences and is gener-
ing what is happening in the organization and ally aimed at optimizing the performance of a
understanding some underlying trends and causes system. The goal here is to provide a decision or
of such occurrences. This involves consolida- a recommendation for a specific action. These
tion of data sources and availability of all rel- recommendations can be in the forms of a spe-
evant data in a form that enables appropriate cific yes/no decision for a problem, a specific
reporting and analysis. From this data infras- amount (say, price for a specific item or airfare
tructure, we can develop appropriate reports, to charge), or even a complete set of produc-
queries, alerts, and trends using various report- tion plans. The decisions may be presented to
ing tools and techniques. In our view, this has a decision maker in a report, or may directly
been the key focus of business intelligence (BI). be used in an automated decision rules system
A significant technology that has become a key (e.g., in airline pricing systems). Thus, these
player in this area is visualization. Using the types of analytics can also be termed Decision
latest visualization tools in the market, we can or Normative Analytics.
now develop powerful insights into the opera-
tions of an organization. Over the years, the use of the analytics and its
related fields has grown tremendously and now
Predictive analytics aims to determine what is there are wide varieties of industry participants
likely to happen in the future. This analysis is that use analytics. They range from industries
based on statistical techniques as well as other that provide data infrastructure, data warehous-
more recently developed techniques that fall un- ing solutions, middleware, data aggregation,
der the general category of data mining. The and analytics software developers to analytics
goal of these techniques is to be able to predict user organizations and academics. Research
if the customer is likely to switch to a com- across all three categories of analytics also con-
petitor (“churn”), what a customer is likely to tinues to flourish. Technology enhancements
buy next and in what quantity, what promo- have enabled application of analytic techniques
tion a customer would respond to, whether this to gain insights from large amounts of both
Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples 173

structured and unstructured data. This forms Many other organizations exist in this industry.
the basis of Big Data analytics. Our goal is not to create a directory of players or
This paper is organized as follows. In Sec- their capabilities in each space, but to illustrate
tion 2, we identify various participants in the to students considering a career in analytics that
analytics industry and group them into clusters. many different options exist for playing in the
Section 3 discusses research in predictive ana- analytics industry. One can start in one sector,
lytics and introduces an example of a predictive and move to another role altogether. We will
analytics project the authors have worked on also see that many companies play in multiple
recently. Section 4 presents a discussion on sectors within the analytics industry and thus
research options in Big Data analytics. This offer opportunities for career transitions within
section also includes a brief description of one the field both horizontally and vertically.
of our research projects that employs Big Data Figure 2 presents the analytic industry clus-
analytics. Lastly, we discuss the role of aca- ters categorized into nine key sectors or clus-
demic programs and modules in advancing the ters, forming Analytics Ecosystem. The first
cause of both research and practitioner devel- five clusters can be broadly termed technology
opment of business analytics. We note that this providers or accelerators. Their primary rev-
paper is an extension of a paper presented at enue comes from developing technology, solu-
the 35th International Conference on Informa- tions, and training which enables the user or-
tion Technology Interfaces, Dubrovnik, Croatia, ganizations to employ these technologies in the
June 2013 [18]. most effective and efficient manner. The accel-
erators include academics and industry organi-
zations whose goal is to assist both technology
2. Analytics Industry Participant Clusters providers and users.

This section is aimed at identifying various an-

alytics industry players by grouping them into 2.1. Data Infrastructure Providers
various sectors. We note that the list of company
names included is not exhaustive. These merely This cluster includes all major organizations
reflect our own awareness and mapping of com- that provide the hardware and software infra-
panies’ offerings in the analytics industry space. structures targeting the basic foundation for
Additionally, the mention of a company’s name the data management solutions. Some of the
or its capability in one specific group does not most obvious players in this cluster: companies
mean that the specific activity is the only offer- providing hardware infrastructure for database
ing of that organization. We use these names computing like IBM, Dell, HP and Oracle; stor-
simply to illustrate our descriptions of sectors. age solution providers like EMC and NetApp;

Figure 2. Analytics industry clusters.

174 Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples

companies providing indigenous hardware and include: Nielsen, which provides data sources
software platforms such as IBM and Teradata; to their clients on customer retail purchase be-
data solution providers offering hardware and havior; Experian, which provides data on each
platform independent database management sys- household in the US; Omniture, which has de-
tems like the SQL Server family of Microsoft; veloped technology to collect web clicks and
specialized integrated software providers such share such data with their clients; and Google,
as SAP; companies like Amazon and Sales- which compiles data for individual websites and
force.com providing full data storage solutions makes a summary available through Google An-
through cloud computing; the recent crop of alytics services.
companies in big data space like Cloudera, Hor-
tonworks, Alteryx, Pivotal and Hadapt; and
companies providing infrastructure services and 2.5. Analytics Focused Software
training to create Big Data platform such as Developers
Hadoop Clusters, MapReduce, NOSQL.
Companies in this category have developed an-
alytics software for general use with data that
2.2. Data Warehouse Industry has been collected in a data warehouse or is
available through one of the platforms identi-
Companies with a data warehousing focus pro- fied earlier (including Big Data). It can also
vide technology and services aimed towards in- include inventors and researchers in universi-
tegrating data from multiple sources thus en- ties and other organizations that have developed
abling organizations to derive and deliver value algorithms for specific types of analytics appli-
from its data assets. They form the backbone for cations. We can identify major industry players
all the players in the analytics industry. Many in this space along with the three types of ana-
companies in this space include their own hard- lytics outlined earlier.
ware to provide efficient data storage, retrieval,
and processing. Recent developments in this
space include performing analytics on the data 2.5.1. Reporting/Analytics
directly in memory. Companies such as IBM,
Oracle, and Teradata are major players in this
Reporting or descriptive analytics is enabled by
arena.
the tools available from the Middleware indus-
try players or unique capabilities offered by
2.3. BI Platforms/Middleware Industry focused providers. For example, Microsoft’s
SQL Server BI tool kit includes reporting as
The general goal of middleware industry is to well as predictive analytics capabilities. On
provide easy to use tools enabling reporting the other hand, specialized software is available
or descriptive analytics on the enterprise-wide from companies such as Tableau for visualiza-
data. This forms a core part of BI or an- tion. SAS also offers a Visual Analytics tool
alytics employed at organizations. Examples for similar capacity. Likewise, there are many
of companies in this space include Microstrat- open source visualization tools as well.
egy, Plum, Oracle–Hyperion, SAP–Business
Objects, IBM-Cognos and many others.
2.5.2. Predictive Analytics

2.4. Data Aggregators/Distributors Perhaps the biggest recent growth in analytics

has been in this category, resulting in a large
These companies typically focus on a specific number of companies that focus on providing
industry sector for specialized data collection, both technical and management solutions for
aggregation and distribution through their niche predictive analytics. Many statistical software
platforms and services for data collection. They companies such as SAS, SPSS, etc. embraced
build their relationships by providing the data to predictive analytics early on and developed the
other industry players and enabling them to em- software capabilities as well as industry prac-
ploy analytics. Some of the example companies tices to employ data mining techniques as well
Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples 175

as classical statistical techniques for analytics. able from the data infrastructure, data ware-
IBM-SPSS Modeler from IBM and Enterprise house, middleware, data aggregators, and an-
Miner from SAS are some of the examples of alytics software providers to develop custom
tools used for predictive analytics. Other play- solutions for a specific industry. Thus, this
ers in this space include KXEN, Statsoft, Sal- industry group makes it possible for the ana-
ford Systems, and scores of other companies lytics technology to be truly useful. Most major
that may sell their software broadly or use it analytics technology providers like IBM, SAS,
for their own consulting practices. Two open
source platforms (R and RapidMiner) have also Teradata, etc., clearly recognize the opportunity
emerged as popular industrial strength software to connect to a specific industry or client and of-
tools for predictive analytics and have compa- fer analytic consulting services. Companies that
nies that support training and implementation have traditionally provided application/data so-
of these open sources tools. lutions to specific sectors are now developing
industry specific analytics offerings. For exam-
ple, Cerner provides electronic medical records
2.5.3. Prescriptive Analytics (EMR) solutions to medical providers and their
offerings now include many analytics reports
Software providers in this category offer model-
and visualizations. Similarly, IBM offers a
ing tools and algorithms for optimization of op-
erations tagged as management science/opera- fraud detection engine for the health insurance
tions research (MS/OR) software. This field industry, and is working with an insurance com-
has had its own set of major software providers. pany to employ their famous Watson analytics
IBM, for example, has classic linear and mixed platform in assisting medical providers and in-
integer programming software, IBM–ILOG, surance companies with diagnosis and disease
which provides prescriptive analysis services. management. Another example of a vertical ap-
Analytics providers such as SAS have their own plication provider is Sabre Technologies, which
OR/MS tools – SAS/OR. Other major players provides analytical solutions to the travel indus-
in this domain include companies such as AI- try including fare pricing for revenue optimiza-
IMS, AMPL, Frontline, GAMS, Gurobi, Lindo tion, dispatch planning, etc.
Systems, Maximal, and many others.
Of course, there are many techniques that fall This cluster also includes companies that have
under the category of prescriptive analytics, developed their own domain specific analyt-
each having their own set of providers. For ex- ics solutions and market them broadly to a
ample, simulation software is provided by major client base. For example, Axiom has devel-
companies like Rockwell (ARENA) and Simio. oped clusters for virtually all households in
Similarly, Frontline offers tools for optimization
the US based upon all the data they collect
with Excel spreadsheets. Decision analysis in
multi-objective settings can be performed us- about households from many different sources.
ing tools such as Expert Choice. There are also Credit score and classification reporting com-
tools from companies such as Exsys, XpertRule, panies (Equifax, Experian, TransUnion, etc.)
and others for generating rules directly from also belong to this group. Demandtec (now
data or expert inputs. Some new companies are owned by IBM) provides pricing optimization
evolving to combine multiple analytics models solutions in retail industry. This field represents
in the Big Data space. For example, Teradata an entrepreneurial opportunity to develop in-
Aster includes its own predictive and prescrip- dustry specific applications. Examples of such
tive analytics capabilities in processing big data companies and their activities are: Sense Net-
streams. works, which employs location data for devel-
oping user/group profiles, X+1 and Rapleaf,
2.6. Application Developers: Industry which profile users on the basis of email usage,
Specific or General Bluecava, which aims to identify users through
all device usage, and Simulmedia, which targets
The organizations in this group use their analyt- advertisements on TV on the basis of analysis
ical expertise, focusing on using solutions avail- of a user’s TV watching habits.
176 Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples

2.7. Analytics User Organizations Support Systems (SIGDSS), a subgroup of the

Association for Information Systems, also fo-
Clearly, this is the economic engine of the whole cuses on analytics. Most of the major vendors
analytics industry. If there were no users, there (e.g., Teradata, SAS, etc.) also have their own
would be no analytics industry. Organizations membership-based user groups promoting the
in every other industry, size, shape, and location use of analytics.
are using analytics or exploring use of analytics A third group of analytics industry analysts
in their operations, trying to gain/retain a com- is what we call analytics ambassadors, influ-
petitive advantage. These include organizations
encers, or evangelists. They have presented
in private sector, government, education, mili-
their enthusiasm for analytics through their sem-
tary, etc. Examples of uses of analytics in dif-
inars, books, and other publications. Illustrative
ferent industries are too numerous to list here.
examples include Steve Baker, Tom Davenport,
Also, specific companies are not identified in
Charles Duhigg, Wayne Eckerson, Bill Franks,
this section. Rather, the goal is to see what type
Malcolm Gladwell, Claudia Imhoff, Bill Inman,
of roles analytics professionals can play within
and many others.
a user organization.
Of course, the top leadership of an organization,
(Chief Information Officer, etc) is critically im- 2.9. Academic Providers and Certification
portant in applying analytics to its operations. Agencies
Although not enough senior managers seem to
subscribe to this view, the awareness of apply- Academic providers and certificate agencies play
ing analytics within an organization is growing a major role in fueling the analytics industry
everywhere. Major organizations in every in- with the necessary human resources by prepar-
dustry that we are aware of are hiring analytical ing students with current analytics expertise.
professionals under a variety of titles; these an- This cluster, made up of different programs
alytics professionals are eventually moving into in business schools (information systems, man-
management positions and driving business de- agement, marketing etc.), hence represents the
cision making using insights derived by em- academic programs that prepare professionals
ploying analytics. for the analytics industry. Other academic pro-
grams such as computer science, statistics, math-
ematics and industrial engineering offer courses
2.8. Analytics Industry Analysts that help train students for the analytics industry.
and Influencers
The certification programs established by tech-
This cluster of analytics industry includes three nology providers such as IBM, Microstrategy
types of organizations or professionals. The and SAS have a paramount aim of giving stu-
first group is the set of professional organiza- dents the needed set of tools that are relevant
tions that provide advice to the analytics indus- to solving real world problems in the indus-
try providers and users. Their services include try. Accompanying exercises and laboratory
marketing analyses, coverage of new develop- assignments are mostly adapted from actual im-
ments, evaluation of specific technologies, and plementations in businesses. For instance, IN-
development of training and white papers. Ex- FORMS has just introduced a Certified Ana-
amples of such players include organizations lytics Professional (CAP) certificate program
such as the Gartner Group, The Data Warehous- that is aimed at testing an individual’s general
ing Institute, Forrester, McKinsey, and many of analytics competency.
the general and technical publications and web- Collaborative initiatives also exist between for-
sites that cover the analytics industry. mal academic institutions and technology provi-
The second group includes professional soci- ders that help students develop strong theoret-
eties or organizations that also provide some of ical background of analytic concepts, sharpen
the same services, but are membership-based. their problem formulation techniques and help
For example, INFORMS, a professional orga- them solve real world problems. Among other
nization, has now focused on promoting an- targets, the collaborative initiatives between aca-
alytics. Special Interest Group on Decision demic institutions and technology providers are
Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples 177

aimed at helping students develop better tools, could result. Using analytics to understand in-
procedures and algorithms through research ac- juries can help in deriving valuable insights that
tivities. The outputs of such research activities would enable the coaches and the medical team
are then translated into providing both rigorous to manage the team composition, understand
and practically sound solutions to real world the player profiles and ultimately aid in better
problems encountered by businesses. decision making on the availability of players.
In an exploratory study, injuries related to Amer-
3. Predictive Analytics Research Examples ican Football were analyzed using reporting and
predictive analytics by our research team. The
project followed the CRISP-DM methodology
Research in applying predictive analytics has to understand the business problem of making
been growing exponentially over the last few recommendations on managing injuries, under-
years. The first wave of the research focused stand the various data elements collected about
on developing algorithms in major categories injuries, cleaning the data, performing relevant
of predictive analytics: classification, cluster- imputations for the missing data, developing vi-
ing, association mining, etc. These included sualizations to draw various inferences, building
decision trees, apriori algorithms, neural net- predictive models to analyze the injury healing
works, and hundreds of other algorithms and time period and drawing sequence rules to pre-
variants. These algorithms have been applied dict the relationship among the injuries and the
in thousands of exploratory projects. A cata- various body parts that were afflicted with in-
log of such research projects will fill a book, juries.
hence we will only illustrate these opportuni-
ties by providing a brief description of one re- The injury data set consisted of over 560 foot-
cent project in applying predictive analytics in ball injury records that include injury-specific
healthcare and sports. variables: body part, body site, laterality, ac-
tion taken, severity, injury type, injury start date
and healing date. It also included player iden-
3.1. Analyzing Athletic Injuries tity and other sport specific variables such as
position played, activity, onset, and game loca-
Any athletic activity is prone to injuries and if tion. Based on the injury start date and the heal-
the injuries are not handled properly, adverse ing date, healing time was calculated for each
performance-related impact to athletic teams record and classified into different sets of time

Figure 3. Injury recovery period at various positions.

178 Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples

periods: 0-1 month, 1-2 months, 2-4 months, cause of injuries in offensive players, such as
4-6 months and 6-24 months. The overall in- the running back and wide receiver positions.
jury records with valid recovery time limited the
final data available for modeling to 384 records. Our analysis included many more visualizations
that enabled better and richer understanding of
We first began with descriptive analytics. Vari- the problem domain and the data. We then be-
ous visualizations were built to draw inferences, gan to apply predictive modeling methods to
depicting the healing time period associated
be able to understand the injury recovery rates.
with player’s positions, severity of injuries and
IBM SPSS Modeler was used for predicting
the healing time period, treatment offered and
the associated healing time period, major in- each of the healing time period categories. The
juries afflicting body parts, type of field and data was divided into training and validation
game etc. Some of these are shown in the fig- data sets and several predictive models using
ures below. techniques like CHAID, CRT and Neural Net-
works were built. Of all the techniques, a neural
Figure 3 depicts the period of recovery from in- network using multi-layer perceptron and a sin-
juries sustained by players at various positions. gle hidden layer with 15 neurons yielded the
The injury recovery variable ranges are 0 to 1 best results.
week, 1 to 2 weeks, 2 to 3 weeks and 3 to 4
weeks. It is realized that fifty percent of in- Some of the predictor variables in the neural
juries are resolved in less than 2 weeks. Also, network model were: current status of injury,
defensive positions are more likely to sustain severity, body part, body site, type of injury, ac-
injuries as compared to other positions. tivity, event location, action taken and position
The visualization in Figure 4 indicates the var- played. The prediction results shown in Table 1
ious motions that lead to players’ injuries. The depict the accuracy of the model in predicting
players are further segmented into different play- how long an athlete remains injured once an in-
ing positions. It is seen from Figure 4 that tack- jury is sustained. The relevant values are the
ling is the main cause of injuries in defensive diagonal figures which indicate the percent ac-
players, such as the defensive secondary and curacy with which the recovery period for dif-
linebacker positions while running is the main ferent periods of injuries could be predicted.

Figure 4. Motions at time of injury to various player positions.

Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples 179

For instance, with an accuracy of 75%, injuries letting the training room staff screen the injured
that take 2 to 4 months to heal can be correctly players; training players at defensive positions
predicted to take that long to heal based on the to increase their reflexive abilities in order to
variables listed above in the prior sections. The avoid getting injuries, and taking precautionary
overall percentage accuracy of the correct pre- measures for the future injuries that a player
dictions of healing time period categories was might suffer based on results of the sequence
89.8%. The classification for the different heal- analysis.
ing periods was generated after initial analysis This research project illustrates an interesting
of the data set and in consultation with industry application of predictive modeling methods in
expects. The relevant injury period of inter- athletic health issues. The next section de-
est with regards to college sports was chosen scribes an emerging area of research in ana-
as a maximum of 2 years since most players lytics.
only start active game time participation in the
last two years of being on the football team.
In this preliminary study, a classification pe- 4. Big Data Research Opportunities
riod in months was chosen in order to give a
broader level view of predicting injury recovery
periods. In order to derive an in-depth under- The bottleneck to knowledge generation and
standing of the injury recovery pattern, further scientific progress is now an issue of effective
data analysis rather than data availability. Con-
studies would drill down and consider injury
temporary lifestyles of individuals and techno-
recovery periods in weeks and days. logical advancements in most parts of the world
have resulted in the generation of huge amounts
of data in astounding proportions every hour, in
Predicted (months)
the form of emails, weblogs, tweets, sensor gen-
Observed erated data, phone logs, GPS data, traffic data,
(month) 0-1 1-2 2-4 4-6 6-24
video, and text. The data sources are a mixture
0-1 97.3% 2.2% 0.5% 0.0% 0.0% of both structured and unstructured data for-
1-2 44.8% 55.2% 0.0% 0.0% 0.0% mats. Hitherto, large volumes of structured data
2-4 18.8% 6.2% 75.0% 0.0% 3.9% have been mainly collected, stored and mined
4-6 16.7% 0.0% 0.0% 83.3% 0.0% using traditional and relational data warehous-
6-24 0.0% 0.0% 25.0% 0.0% 75.0% ing technologies. However, the fast growth of
data has resulted in a new breed of analytics
called Big Data analytics. Big Data analytics
Table 1. Classifications for healing category.
supports large scale data analysis that focuses
on large volumes of historical and near real-time
Furthermore, we also performed a sequence data of different varieties and that needs to be
analysis with injuries and body parts as prece- analyzed at a fast pace.
dents and antecedents respectively. The players At the core of Big Data research are busi-
are most likely to suffer the same injury again ness concepts, technologies and statistical tech-
(repeated injuries) associated with same body niques necessary to design highly scalable sys-
part. It was also interesting to note that lower tems that can collect, process, store and analyze
extremity injuries resulted in upper extremity large volumes of both structured and unstruc-
injuries. Some of these include possible injuries tured data. One of the most popular platforms
afflicting the hand when a player had already used to support processing of Big Data is an
had an injury to the ankle and injuries afflict- open source platform called Apache Hadoop.
ing shoulder when a player has had an injury to Apache Hadoop has two main basic compo-
the femur. For instance, with a support of 25% nents; Hadoop Distributed File System (HDFS)
and confidence of 22.5%, a strain to the femur and MapReduce. HDFS, a distributed and scal-
would result in a consequent shoulder injury. able file system contributed by Google, supports
the storage of data and computations on multi-
Based on the analysis, many business recom- ple commodity computers. MapReduce is a
mendations were suggested which include the high level programming model that allows for
following: employing the services of healthcare parallel and distributed algorithms to be writ-
specialist at the onset of the injury instead of ten for processing massive sets of data on a
180 Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples

cluster of computers. Besides these two com- such as physicians, nurses and other clinical ex-
ponents, the open source community has de- perts, the advent of the Web 2.0 platform has
veloped other sub-projects that help efficiently ushered an era where patients are beginning to
manage massive amounts of data. Some of inculcate new healthcare delivery strategies in
these sub-projects are Hive, Pig, Mahout and the self-management of chronic diseases. In
Hbase. For instance, Hive is an SQL-like data fact, some studies have suggested that tradi-
warehouse platform that supports data opera- tional sources of health management informa-
tions such as data summarization and complex tion may no longer be enough in providing qual-
queries of data sets. The sub-projects may have ity healthcare to patients [4]. Pharmaceutical
different or over-lapping roles, yet they all are companies and other healthcare-related busi-
designed to facilitate efficient storage, manipu- nesses already market drugs and solicit con-
lation and analysis of massive amounts of data. sumer feedback from social networks about new
The Apache Hadoop platform is not a replace- drugs and disease management options.
ment for traditional relational database systems,
but rather a complement that can better handle Among a myriad of health related topics dis-
both structured and unstructured data sources. cussed on social media platforms is the man-
agement of chronic diseases such as diabetes,
Just like other streams in the data analytics field, chronic obstructive pulmonary disease (COPD),
research in Big Data analytics is applicable to major depressive disorder (MDD), epilepsy and
almost all fields of practice and study. In areas cardiac failure [2]. These social media plat-
like healthcare, logistics, marketing, retail and forms offer social support and information to
defense, it has become imperative to explore patients and caregivers. Social support has been
and analyze large amount of data in ways that identified as a major factor for improving health
supports effective decision making and strate- outcomes in patients who suffer from various
gic management. Such research and practice types of illness [3][6]. The concept of social
is growing rapidly. The highly scalable plat- support, also sometimes referred to as peer sup-
form used for Big Data analytics is necessi-
port has been shown to not only help disease
tated by three fundamental characteristics of
prevention, but also foster and promote general
data sources; volume, velocity and variety. Big
Data research looks at how to analyze data in well-being and health [15][14][7]. The concept
different domains with such characteristics and of social relationships and how they can be es-
in a way that generates deeper knowledge and tablished on electronic social media platforms
adds value to the decision making process in in the context of the provision of healthcare
businesses. In the following, we describe one support is still at an early stage in academic
project on which we are currently working. research. More research needs to be conducted
to establish the role and effectiveness of social
media platforms in healthcare delivery.
4.1. Analyzing the Role of Social Media in Previous studies have mostly investigated ef-
Healthcare Delivery forts made by healthcare institutions in using
social media as a tool for healthcare manage-
One of the nascent research streams in Big ment and information delivery [9]. Armstrong
Data analytics is the role of electronic social and Powell (2009) assert that patients value
media platforms in healthcare delivery. Early the additional support they receive from on-
studies have established that factors such as line communities and perceive online sources
evolving patient demographics, financial con- of information provided by lay people as a
straints and longer incidence of chronic diseases complement to information provided by pro-
have enhanced the impact of social relationships fessional healthcare providers [1]. In addition,
in healthcare delivery [15]. Social relation- studies have shown that peer-to-peer support
ships, which are associative bonds established creates an enabling environment for disease
between individuals with common aspirations self-management. However, very few studies
and for the purpose of giving and receiving sup- have investigated how disease management and
port, foster disease self-management programs healthcare support are initiated by patients in
that are vital to the management of chronic dis- a social media domain. Patients like to be in
eases. contact with others with similar problems in or-
Even though healthcare delivery and support der to draw both emotional support and obtain
has been traditionally made through channels information on how to manage diseases [1].
Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples 181

It has been established that social media plat- 5. Teaching Perspectives

forms are a viable avenue for disease self-mana-
gement. The question, however, still remains
In any knowledge intensive industry such as an-
whether self-management programs through so- alytics, the fundamental strength comes from
cial media platforms are an effective means having students who are interested in the tech-
of managing chronic diseases. Disease self- nology and choose that industry as their pro-
management, especially for chronic diseases, fession. Universities play a key role in mak-
can be an important aspect of the healthcare de- ing this possible. Academic programs that
livery process. Self-management programs are prepare professionals for the industry are in-
designed as a complement to other treatment creasing in number and type. These programs
methods to put the patients in the driving seat span various components of business schools
and empower them to take control of their health such as information systems, marketing, man-
care status. The aim of a self-help program in- agement sciences, etc. Such programs are also
cludes emerging far beyond business schools to include
computer science, statistics, mathematics, and
1. teaching the appropriate use of medications, industrial engineering departments across the
2. evaluating methods and use of new treatment world. These programs also include training of
methods, graphics developers who design new ways of vi-
sualizing information. Universities are offering
3. engaging in physical exercises that help de- undergraduate and graduate programs in analyt-
crease instance of fatigue, isolation and pain, ics in all of these disciplines, though they may
be labeled differently. A major growth fron-
4. better ways of communicating with family, tier has been certificate programs in analytics
peers and clinical healthcare givers. to enable current professionals to re-train and
re-tool themselves for analytics careers. Cer-
There is a need to investigate the effect of peer tificate programs enable practicing analysts to
support and information exchange on social me- gain basic proficiency in specific software by
dia in self-managing chronic disease conditions taking a few critical courses. Power (2012)
in an environment that is not directly mediated published a partial list of the graduate programs
by a healthcare institution or any clinical stake- in analytics, but there are likely many more such
holder. Sarasohn-Kahn (2008) reported that, programs, with new ones being added daily [11].
with respect to quality, expert information pro- Watson (2013) describes how he teaches his an-
vided by a large collection of users on a social alytics courses and also argues for the growing
media platform is known to be at par with what popularity of such programs [17]. Graduates
would be provided by any single expert on a from these programs fill such positions as data
subject matter [12]. The same is true for clin- analysts, data scientists and data mining archi-
ical information provided by other users on a tects.
social networking platform. Hence, besides the
healthcare advice from expert clinicians such as Another group of players assist in developing
doctors, the advice and support provided on so- competency in the field of analytics. These
cial media platforms by non-expert patients is are certification programs awarding certificates
of expertise in specific software. Most (if
useful to the achieving of the necessary health
not all) major technology providers (IBM, Mi-
goals that need to be attained by patients. crosoft, Microstrategy, Oracle, SAS, Teradata,
As the focus of traditional health informatics etc.) have their own certification program.
shifts from health professionals to consumers, These certificates ensure that potential new hires
Big Data research assumes a vital role because it have a certain level of tool skills. The Certified
gives one the opportunity and flexibility to ana- Analytics Professional (CAP) certificate pro-
lyze extremely huge amounts of data rather than gram introduced by INFROMS has the main
just a small sample and make inferences and aim of equipping individuals with a general
decisions on the use and delivery of healthcare competency in data analytics. Any of these
on social media platforms. Big Data research certifications gives a college student additional
also delves into in-depth analysis and drawing marketable skills.
of deeper conclusions on data generated from Many of these industry leaders also have aca-
other domains of study. demic alliance programs to enable teaching their
182 Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples

software and its uses. IBM, Microsoft, Ora- 6. Conclusion

cle, SAP, and Teradata have major initiatives
that allow faculty members to use and share Overall, there is much to be excited about in the
software as well as other teaching materials analytics industry at this point. Its applications
for instructional purposes. Watson (2013) pro- stretch to almost all fields of practice and re-
vides a summary of several of these resources search including healthcare, retail and manufac-
[17]. For example, Teradata University Network turing. Support for teaching analytics is avail-
(www.teradatauniversitynetwork.com) also able from many major vendors. The industry
known as TUN is sponsored and supported by clusters of analytics ecosystem show that there
Teradata, a major provider of analytics hard- are many opportunities for students, academics
ware/software and applications consulting ser- and industry professionals. With increases in
data availability and the realization that more
vices. TUN is free to use and provides faculty business insights could be generated from an
members and students access to many case stud- otherwise junk volume of data, research in Big
ies, white papers, teaching materials, and soft- Data has increased. This trend in Big Data
ware for learning concepts, applications, and analytics is expected to continue to grow and
tools for analytics. Faculty members contribute complement traditional data analytics methods
such resources to TUN. Even teaching notes and existing data management strategies. The
are available for many of the teaching materi- increase in the amount of attention analytics is
als. A recent addition is several videos that de- receiving from the press also indicates that this
pict analytics applications in realistic settings. is an expanding field which can provide a com-
These videos have been developed by Dr David petitive edge to many types of businesses. As
illustrated in this paper, descriptive, predictive
Schrader of Teradata. These are labeled Busi- and prescriptive analytics applications and re-
ness Scenario Investigations (BSI), a takeoff on search opportunities are wide and growing.
the popular TV show CSI. Not only are these en-
tertaining, but they also provide the class some
questions for discussion. For example, in at References
http://www.teradatauniversitynetwork.
com/teach-and-learn/library-item/?
LibraryItemId=889, one can assume the role [1] N. ARMSTRONG, J. POWEL, Patient perspectives on
health advice posted on Internet discussion boards:
of an airline customer service center profes- a qualitative study. Health Expectations, 12(13)
sional. A video on YouTube gives informa- (2009), 313–320.
tion about customers’ profiles and relationship [2] G. BOND, R. BURR, F. WOLF, M. PRICE, S. MC-
with the airline. An incoming flight is running CURRY, L. TERI, The Effects of a Web-Based Inter-
late, and several passengers are likely to miss vention on the Physical Outcomes Associated with
Diabetes among Adults Aged 60 and Older: A
their connecting flights. There are seats on one Randomized Trial. Diabetes Technology and Ther-
outgoing flight that can accommodate two of apeutics, 9(1) (2007), 52–59.
the four passengers. Students watch the video, [3] C.-L. DENNIS, Peer support for postpartum depres-
pause it as appropriate, and answer the questions sion: volunteers’ perceptions, recruitment strate-
on which passengers should be given priority. gies and training from a randomized controlled
They then resume the video to get more infor- trial. Health Promotion International, 28(2) (2013),
187–196.
mation. Students’ decisions might change as
they learn more about those customers’ profiles. [4] M. Evans, L. Donelle, L. Hume-Loveland, Social
support and online postpartum depression discus-
TUN also provides information about how the sion groups: A content analysis. Patient Education
analysis was prepared in a link at the bottom of and Counseling, 87(3) (2012), 405–410.
the video page. This engaging multimedia ex- [5] INFORMS, Analytics Section Overview.
cursion provides an example of how additional http://www.informs.org/Community/
information available through an enterprise data Analytics, Accessed February 2013.
warehouse can assist in decision making. There [6] B. LAKEY, S. COHEN, Social support theory and
are almost a dozen such videos and hundreds measurement. In S. COHEN, L. G. UNDERWOOD,
B. GOTTLIEB (EDS.), Social Support Measurement
of other teaching resources available through and Intervention: A Guide for Health and Social
TUN. Scientists. Oxford University Press, (2000) Toronto.
Research and Pedagogy in Business Analytics: Opportunities and Illustrative Examples 183

[7] M. E. LARA, J. LEADER, D. N. KLEIN, The associ- Received: August, 2013

ation between social support and course of depres- Revised: September, 2013
sion: Is it confounded with personality? Journal of Accepted: September, 2013
Abnormal Psychology, 106 (1997), 478–482.
Contact addresses:
[8] E. K. LEE, F. PIETZ, B. BENECKE, J. MASON, G. Ramesh Sharda
BUREL, Advancing Public Health and Medical Pre- Institute for Research in Information Systems
paredness with Operations Research. Interfaces, Spears School of Business
43(1) (2013), 79–98. Oklahoma State University
Stillwater, OK 74078
[9] M. P. MARTINASEK, A. D. PANZERA, T. SCHNEI- USA
DER, J. H. LINDENBERGER, C. A. BRYANT, R. J. e-mail: [email protected]
MCDERMOTT, M. COULURIS, Benefits and Barriers Daniel Adomako Asamoah
of Pediatric Healthcare Providers toward Using So- Institute for Research in Information Systems
cial Media in Asthma Care. American Journal for Spears School of Business
Health Education, 42(4) (2011). Oklahoma State University
Stillwater, OK 74078
[10] P. PEKGUN, R. P. MENICH, S. ACHARYA, P. G. USA
FINCH, F. DESCHAMPS, K. MALLERY, J. V. SISTINE, e-mail: [email protected]
K. CHRISTIANSON, F. J. CARLSON REZIDOR, Ho- Natraj Ponna
tel Group Maximizes Revenue Through Improved Institute for Research in Information Systems
Demand Management and Price Optimization. In- Spears School of Business
terfaces, 43(1) (2013), 21–36. Oklahoma State University
Stillwater, OK 74078
[11] D. P. POWER, What universities offer master’s USA
degrees in analytics and data science. (2012). e-mail: [email protected]
http://dssresources.com/faq/index.php?
action=artikel&id=250, Accessed Feb 2013.
[12] J. SARASOHN-KAHN, The wisdom of patients: RAMESH SHARDA is Director of the PhD in Business for Executives
Health care meets online social media. Oakland, Program, the Institute for Research in Information Systems (IRIS),
CA: California HealthCare Foundation. (2008). ConocoPhillips Chair and Regents Professor of management science
http://www.chcf.org/publications/2008/ and information systems in the Spears School of Business at Oklahoma
04/the-wisdom-of-patients-health-care- State University. He has co-authored two text books (Decision Sup-
port and Business Intelligence Systems, 9th edition, Prentice Hall and
meets-online-social-media, Accessed Oct Business Intelligence: A Managerial Approach, 2nd Edition, Prentice
2012. Hall). His research has been published in major journals in manage-
ment science and information systems including Management Science,
[13] SAS.COM, Eight Levels of Analytics Operations Research, Information Systems Research, Decision Support
http://www.sas.com/news/sascom/ Systems, Interfaces, INFORMS Journal on Computing, and many oth-
analytics levels.pdf, Accessed Feb 2013. ers. He is a member of the editorial boards of journals such as the
Decision Support Systems, ACM Transactions of MIS, and Information
[14] I. SKARSATER, A. LANGIUS, H. AGREN, L. HAG- Systems Frontiers. He is currently serving as the Executive Director of
GSTROM, K. DENCKER, Sense of coherence and so- Teradata University Network. He recently received the 2013 INFORMS
cial support in relation to recovery in first-episode HG Computing Society Lifetime Service Award.
patients with major depression: A one-year prospec-
tive study. International Journal of Mental Health
Nursing, 14 (2005), 258–264. DANIEL ADOMAKO ASAMOAH is pursuing a PhD in management sci-
ence and information systems in the Spears School of Business at the
[15] M. STEWART, V. TILDEN, The contributions of health Oklahoma State University. He obtained a Master of Science degree
care science to social support. International Journal in telecommunications management from the same university. Prior
of Nursing Studies, 32 (2005), 535–544. to his graduate studies, he completed a Bachelor of Science degree in
electrical and electronic engineering at the Kwame Nkrumah University
[16] T. VARELAS, S. ARCHONTAKI, J. DIMOTIKALIS, O. of Science and Technology, Ghana. His research has been published
TURAN, I. LAZAKIS, O. VARELAS, Optimizing Ship in journals such as Decision Support Systems and Simulation: Trans-
Routing to Maximize Fleet Revenue at Danaos. actions of the Society for Modeling and Simulation International. His
Interface, 43(1) (2013), 37–47. current research interests include the use of Big Data analytics as a
platform to study the role of social media in healthcare provision. His
[17] H. WATSON, The Business Case for Analytics. research also involves business intelligence, data analytics, decisions
BizEd, May/June (2013), 49–54. support systems in healthcare and the use of analytical modeling in
solving operations management and information systems problems.
[18] R. SHARDA, D. ASAMOAH, N. PONNA, Business
Analytics: Research and Teaching Perspectives,
Proceedings of the 35th International Conference NATRAJ PONNA is a Master’s student in management information sys-
on Information Technology Interfaces, Dubrovnik, tems at Oklahoma State University. He is a BASE SAS 9 certified
Croatia, June 2013. programmer and a certified SAS predictive modeler with experience
in using SAS Enterprise Miner 7. He also has the SAS and OSU
Data Mining Certificate and has undertaken several projects in business
intelligence and data mining.