Big Data Analytics Services for Enhancing Business Intellige
Introduction
This is the era of big data [1]. Big data and big data analytics have been revolutionizing innovation, research, development as well as management and business [2, 3, 4]. Big data analytics services have created big market opportunities. For example, the researcher of International Data Corporation (IDC) forecasts that big data and analytics-related services marketing in Asia/ Pacific (Excluding Japan) region will grow from US$3.8 billion in 2016 to US$7.0 billion in 2019 at a 16.3% compound annual growth rate (CAGR) [5]. Big data and its emerging technologies including big data analytics have been not only revolutionizing the way the business operates but also making traditional data analy- tics and business analytics bring new big opportunities for acade- mia and enterprises [1, 4, 6, 7]. Big data analytics is an emerging big data technology, and has become a mainstream market adopted broadly across industries, organizations, and geographic regions and among individuals to facilitate big data-driven deci- sion making for businesses and individuals to achieve desired business outcomes [8, 9, 10].
Business intelligence (BI) has received increasing attention in academia, business, and management since 1989 [11], although it was first introduced by an IBM researcher H.P. Luhn in 1958 [12]. BI has become not only an important technology for improving business performance of enterprises but also a marketing brand for developing business, e-commerce, e-services [13]. It is also the momentum for developing organization intelligence, enterprise intelligence, management intelligence, and marketing intelligence [14]. However, what does the intelligence mean in BI? This is still an issue for understanding BI completely. Furthermore, BI is facing new big challenges because of dramatic development of big data and big data technologies [14, 15]; that is, how to use big data analytics services to enhance BI becomes a big issue for business, e-commerce, e-services, and information systems [7].
The article extends the ontology of big data analytics intro- duced in our early work through adding a new level, a technolo- gical level of big data analytics. This article then looks at the temporality, expectability, and relativity of intelligence and con- siders them as the three characteristics of intelligence in BI. Temporality, expectability, and relativity of intelligence are a kind of human intelligence related to business. This article also examines big data analytics as a technology for enhancing BI through examining the relationship between big data analytics and BI. We then review a big data analytics service-oriented architecture (BASOA), in which we also explore how to apply big data analytics services to enhance BI, where we show that the proposed BASOA is viable for enhancing BI based on our surveyed data analysis.
The remainder of this article is organized as follows. “An ontology of big data analytics” extends an ontology of big data analytics introduced in [7]. “Temporality, expectability, and rela- tivity of intelligence in BI” explores the temporality, expectability, and relativity of intelligence in BI. It also looks at BI and its relationships with big data analytics. “BASOA: Big data analytics service-oriented architecture” presents BASOA, a big data analy- tics services-oriented architecture. “Applying BASOA to enhance BI” applies proposed BASOA to BI. The final sections discuss the related work and end this article with some concluding remarks and future work.
An ontology of big data analytics
This section proposes an ontology of big data analytics and looks at the interrelationship between big data analytics and data analytics. To begin with, this section first examines the fundamental of big data analytics.
Big data analytics can be defined as the process of collect- ing, organizing, and analyzing big data to discover, visualize,
and display patterns, knowledge, and intelligence as well as other information within the big data [7, 14]. Similarly, big data analytics can be defined as techniques used to analyze, acquire, and visualize knowledge and intelligence from big data [14]. Big data analytics is an emerging science and technology involving the multidisciplinary state-of-the-art information and communication technology (ICT), mathe- matics, operations research (OR), machine learning (ML), and decision sciences for big data [2, 6]. The main compo- nents of big data analytics include big data descriptive analy- tics, big data predictive analytics, and big data prescriptive analytics [7, 16]. In other words, big data analytics can be represented as
Big data analytics ¼ big data descriptive analytics
þ big data predictive analytics
þ big data prescriptive analytics (1)
where + can be explained as “and.” Equation (1) indicates that big data analytics consists of big data descriptive analytics, big data predictive analytics, and big data prescriptive analytics, also see Figure 1.
● Big data descriptive analytics is descriptive analytics for big data [7, 17, 18], and is used to discover new, nontrivial information [18, p. 2], and explain the characteristics of entities and relationships among entities within the exist- ing big data [19, p. 611]. It addresses the problems such as what happened, and when, as well as what is happening. For example, web analytics for pay-per-click or e-mail marketing data belongs to big data descriptive analytics [20].
● Big data predictive analytics is predictive analytics for big data [7, 16], which focuses on forecasting trends by addres- sing the problems such as what will happen, what’s going to happen, what is likely to happen, and why it will happen [4, 17, 21]. Big data predictive analytics is used to create models to predict future outcomes or events based on the existing big data [19, p. 611]. For example, big data pre- dictive analytics can be used to predict where might be the next attack target of terrorists.
● Big data prescriptive analytics is prescriptive analytics for big data [7, 16], which addresses the problems such as what we should do, why we should do, and what should happen with the best outcome under uncertainty [16, p. 5]. For example, big data prescriptive analytics can be used to provide an optimal marketing strategy for an e-commerce company.
From the above analysis, big data descriptive analytics and big data predictive analytics are the solutions to the challenge of big data to the existing descriptive analytics and predictive analytics, respectively, in turn, to the existing descriptive data mining and predictive data mining, respectively [18, p. 2].
An ontology is a formal naming and definition of a num- ber of concepts and their interrelationships that really or fundamentally exist for a particular domain of discourse [7, 21]. Then, an ontology of big data analytics is an investigation into a number of concepts and their interrelationships that fundamentally exist for big data analytics, as illustrated in Figure 1, based on our early work [7]. In this ontology, big data analytics is at the top while big data and data analytics are at the bottom. Big data descriptive analytics, big data predictive analytics, and big data prescriptive analytics are below the big data analytics as its components.
DW + DM + SM + ML+ Visualization+ Optimization are above Big data and data analytics, where DW, DM, SM, and ML are abbreviations of data warehouse, data mining, statistical modeling, and machine learning, respectively [7]. The current leading DW includes Amazon’s Redshift, Google’s BigQuery, Microsoft’s Azure SQL Data Warehouse, and Teradata [3].
In Figure 1, data analytics refers to as a method or technique that uses data, information, and knowledge to learn, describe, and predict something [21, p. 341]. In brief, data analytics can then be considered as data-driven discoveries of knowledge, intelligence, and communications [17]. More generally, data analytics is a science and technology about examining, summarizing, and drawing conclusions from data to learn, describe, predict, and visualize something [6, 7].
The fundamentals of big data analytics consist of mathematics, statistics, engineering, human interface, computer science, and information technology [2, 6]. The techniques for big data analy- tics encompass a wide range of mathematical, statistical, and modeling techniques [19, p. 590]. Big data analytics always involves historical or current data (often related to operations) and visualization [23]. This requires big data analytics to use DM to discover knowledge from a DW or a big dataset in order to support decision-making, in particular in the text of big business and management [21, p. 344]. DM employs advanced statistical tools to analyze the big data available through DWs and other sources to identify possible relationships, patterns, and anomalies and discover information or knowledge for business decision- making [18, 19]. DW extracts or obtains its data from operational databases as well as from external open sources, providing a more comprehensive data pool including historical or current data [19,
p. 590]. Big data analytics also uses statistical modeling (SM) to discover knowledge and wisdom through descriptive analysis that can support decision-making [7]. Visualization technologies including display technologies as an important part of big data analytics make knowledge patterns and information for decision- making in a form of figure or table or multimedia. In summary, big data analytics in general and big descriptive data analytics, big predictive data analytics, big prescriptive data analytics in specific can facilitate business decision-making and realization of business objectives through analyzing current problems and future trends, creating predictive models to forecast future opportunities and threats, and analyzing/optimizing business processes based on involved historical or current data to enhance organizational performance using the mentioned techniques [17]. Therefore, big data analytics can be represented as shown below [7]:
Big data analytics ¼ Big data þ data analytics þ DW
þ DM þ SM þ ML
þ Visualization þ optimization (2)
Equation (2) reveals the fundamental relationship between big data, data analytics, and big data analytics, that is, big data analytics is based on big data and data analytics, as illustrated in Figure 1. It also shows that computer science and informa- tion technology play a dominant role in the development of big data analytics through providing latest techniques and tools of DM, DW, ML, and visualization [6, 7]. SM and optimization still play a fundamental role in the development of big data analytics, in particular in big data prescriptive analytics [16].
It should be noted that Equation (2) is a concise represen- tation for the technological components of big data analytics whereas the proposed ontology of big data analytics is to look at what big data analytics constitutes at a relatively high level, also see Equation (1). At a relatively lower level, the ontology also illustrates what technologies and techniques can support big data descriptive analytics, big data predictive analytics, and big data prescriptive analytics, as illustrated in Figure 1. Apache Hadoop is a platform of big data analytics [1]. As an open source platform for storing and processing large datasets using clusters and commodity hardware, Hadoop
can scale up to hundreds and even thousands of nodes.
Apache Spark is one of the most popular big data analytics services. It has moved from being a component of the Hadoop system, to the big data analytics platform for a number of enterprises [1, 3]. Spark provides dramatically increased large- scale data processing compared to Hadoop, and a NoSQL database for big data management [1, 19]. Apache Spark has provided Goldman Sachs with excellent big data analytics services [3].
We look at the big data descriptive, predictive, and pre- scriptive analytics as one dimension, and the technological components of big data analytics as another dimension. Then we provide a two-dimensional analysis for big data analytics as a future research work.
Temporality, expectability, and relativity of intelligence in BI
This section explores the temporality, expectability, and rela- tivity of intelligence. It also looks at BI and its relationships with big data analytics.
There are many different definitions on BI from different perspectives. For example,
● BI is a framework that allows a business to transform data into information, information into knowledge, and knowledge into wisdom [19, p. 560]. BI has the potential to positively affect a company’s culture by creating “business wisdom” and distributing it to all users in an organization. This business wisdom empowers users to make sound business decisions based on the accumulated knowledge of the business as reflected on recorded historic operational data [19, p. 560].
● BI refers to as a collection of information systems (IS) and technologies that support managerial decision makers of operational control by providing information on internal and external operations [21].
● BI is defined as providing decision makers with valuable information and knowledge by leveraging a variety of sources of data as well as structured and unstructured information [24].
The first definition of BI emphasizes that BI is a framework and creates business wisdom for decision makers through business data, information and knowledge and their transfor- mations. The second definition stresses “a collection of IS and technologies” while specifies the decision makers to “manage- rial decision makers of operational control,” and information to “information on internal and external operations.” The last definition emphasizes BI “providing decision makers with valuable information and knowledge.” Based on the above analysis, BI can be defined as a framework that consists of a set of theories, methodologies, architectures, systems, and technologies that support business decision-making with valu- able data, information, knowledge, and wisdom. This defini- tion reflects the evolution of BI and its technologies from decision support systems (DSS) and its relations with data warehouses, executive information systems [24].
The principal tools for BI include software for database query and reporting (e.g. SAP ERP, Oracle ERP, etc.), tools for multidimensional data analysis (e.g. OLAP), and DM, e.g. predictive analysis, text mining, web mining [25]. DM is also considered as a foundation of BI [11].
However, what intelligence means in BI is still a big issue for comprehending BI. In what follows, this section addresses this issue through exploring temporality, expectability, and relativity of intelligence in BI.
Terminologically, intelligence in BI should be at least human intelligence in relation to business. This kind of intel- ligence is based on learning and understanding the facts provided by business data, information, and knowledge about a business operation and environment or a product or service [19]. The ability of learning, understanding, and rea- soning belongs to the category of human intelligence [26].
The term “intelligent” has been popular, not only in aca- demia but also in the wider community, due to a long time, ongoing research and development of artificial intelligence (AI) and intelligent systems (IS) since 1955 [27]. There are about 243 million results related to “intelligent” in the Google world (searched on 27 May 2016). In the academia, the term “intelligent” frequently appears in titles of a great number of books, book chapters, papers, and international conferences as well as other media or products. In the wider community, the term “intelligent” often appears in home appliances and cus- tomer electronics including televisions, cameras, vacuum clea- ners, washing machines [7], and mobile phones, to name a few. Defining intelligent is not a simple task. According to the Macmillan Dictionary [28, p. 787], the term intelligence means “the ability to understand and think about things, and to gain and use knowledge.” Similarly, the term intelligence has been defined in IS as “the ability to learn and understand, solve problems and make decision” [29, p. 18]. The term intelligent means to be able to perceive, understand, think, learn, predict, and manipulate a system [27, p. 1]. All these definitions on intelligence are mainly human intelli- gence, which has impacted the development of AI [27]. AI has been focusing on intelligence of machines or machine intelligence (Note that the web is also a machine.). In other words, AI is the science and engineering of making intelligent machines to imitate human intelligence [26]. However, a system may not be considered intelligent, even if it has these abilities associated with human intelligence, because the term intelligent implies some expectations from human beings or society, in particular in the setting of business. Practically, it appears that an intelligent system contains a set of functions that jointly make the system easy to use [30], because “easy” is a term related to human intelligence. More generally, a system or a product is intelligent if and only if it contains a set of functions that jointly make the system either easier or faster, or friendlier, or more efficient, or more satisfactory to use than an existing cognate system taking into account the time. Easier, faster, friendlier, or more efficient, or more satisfactory are all the expectations of humans or customers or society for the performance of a system or product. For example, a high speed train running in China is intelligent, because it is faster and friendlier than the existing ones; these are what the Chinese expect.
The above discussion leads to three perspectives on “intel-
ligence” in BI: temporality of intelligence, expectability of intelligence, and relativity of intelligence.
Temporality of intelligence
There are two meanings for temporal intelligence. (1) Temporal intelligence is the ability to adapt to change. This has been motivated to develop temporal logic and evolutionary comput- ing including genetic algorithms [27]. (2) Temporality of intelli- gence means that intelligence is related or limited to a time interval. For example, at the time of writing this article, few people consider floppy disks as intelligent storage devices. However, a few decades ago floppy disks were considered intel- ligent in comparison to paper tape for data storage. In what follows, we limit ourselves to the meaning of item 2.
Expectability of intelligence
Intelligence can be considered as a substitution for easier, or faster, or friendlier, or more efficient, or more satisfactory. This is expectability of intelligence. We denote them using the degree of satisfaction. All these related concepts are a set of expectations of humans, as parts of human intelligence.
We denote these expectations for a product, P, as EP ei ei is an expected performance for functioni of a product
ei i 1; 2; .. . ; n 1; n , where n is a given integer. For every i 1; 2; .. . ; n 1; n , there is a perceived perfor- mance of customer for functioni, pi, then a product P is intelligent if and only if there exists at least one i 2 f1; 2; ... ; n — 1; ng such that [31, p. 436] s pi >1 (3)
ei
where si is the satisfaction degree of the customer to the ith function of product P.
For example, an iPhone 6S’s Touch ID, Apple’s fingerprint recognition feature, is noticeably quicker when unlocking the phone. “quicker” is what the user perceived, p1 1:5, while “quick” is an expected performance, e1 1; for iPhone 6S from a customer, based on Equation (3), we have the satisfac- tion degree of the customer si 1:5 > 1. Then an iPhone 6S is intelligent.
Relativity of intelligence
Intelligence is a consequence of comparison between two systems (or products or services), which leads to the relativity of intelligence. Generally speaking, let X and Y be two sys- tems. X is intelligent if X is better than Y with respect to E, where E is a set of human expectations. “Better” is a relativity concept. For example, a new microwave is intelligent because it displays the temperature when microwaving food. A user believes that displaying the temperature is better than not displaying it. This example reflects the relativity of intelli- gence. Displaying temperature belongs to the set of expecta- tions E. The set of human expectations can be considered as a set of demands. The expectation of human beings and society promotes intelligence and social development. Therefore, it is significant to define IS with respect to the set of human expectations or demands.
In summary, intelligence in BI can be measured through three dimensions: temporality, expectability, and relativity. In other words, in a BI system there are three characteristics of its intelligence: temporality, expectability, and relativity. The degree of intelligence of a BI system or product or service can be measured using this triad, that is,
Degree of intelligence ¼ temporality þ expectability
þ relativity (4)
Equation (4) is more useful for BI and big data intelligence, because they are based on performance, business advantages, competiveness advantages of systems or products or services. All of these are closely associated with temporality, expect- ability, and relativity of intelligence. This formula can be realized by using big data analytics and big data, in other words, big data and big data analytics can generate big data intelligence, for short,
big data intelligence = big data + big data analytics (5)
Equation (5) indicates that increase of either big data or big data analytics can increase big data intelligence. This is par- tially proved by what Professor Peter Norvig, the Google’s Director of Research, said “we don’t have better algorithms; we just have big data” [4].
Temporality, expectability, and relativity of intelligence can be considered as a fundamental for BI including organization intelligence, enterprise intelligence, marketing intelligence [14], and big data intelligence. We will explore it as a future work.
In fact, the global competitiveness among the giant com- panies lies in these three dimensions of intelligence in busi- nesses, decision-making, products, and systems. Big data and analytics will intensify the competition of the giant companies in terms of temporality, expectability, and relativity of intelli- gence. This might be a reason why giant companies are adopting big data and analytics technologies [3, 32]. This degree of intelligence also differentiates BI from AI using the three properties of intelligence in BI. Throughout this article, we use these three dimensions of intelligence as the basis to understand BI.
Big data analytics can be considered as a part of BI [11, 25], because it “supports business decision making with valuable data, information and knowledge” [6]. Both BI and big data analytics are common in emphasizing either valuable data or information or knowledge. Tableau [3], QlikView, and Tibco’s Spotfire are leading BI tools for interactive visualization for data exploration and discovery [32]. These BI tools are also considered as the tools of big data analytics. This implies that BI and big data analytics share some common tools to support business decision-making.
Currently, BI is based on four cutting-age technology pil- lars: cloud, mobile, big data, and social networking technolo- gies [22, 32] (Note that technology as a service is becoming popular in cloud computing). Each of these pillars corre- sponds to a special kind of web services, that is, cloud services, mobile services, big data services, and social networking ser- vices. All these constitute modern web services, the Internet of services [23]. Each of these services has been supported by big data analytics services [6, 7], as shown in Figure 2. This may be the reason why many companies have been moving from BI, DM, and DW to business analytics in general and big data analytics (BA) in specific in recent years [3, 26].
It should be noted that for the state-of-the-art web services, Sun et al. [7, 23] explore that web services mainly consist of mobile services, analytics services, cloud services, social net- working services, and service as a web service. Here we emphasize big data analytics services at the center to support cloud services, social networking services, mobile services, and e-services to reflect the big data and big data analytics as an emerging new service [7].
Based on IDC’s prediction for the IT market in 2014 [33], spending on big data will explode and grow by 30% to $14+ billion, in which, the spending on big data analytics services will exceed $4.5 billion, growing 21%. The number of providers of big data analytics services will triple in three years. This means that big data analytics services have become an important emerging mar- ket, together with the Internet of services including e-services, cloud services, mobile services, and social networking services. All
these five services and the technologies shape the most important markets for e-commerce and e-business [23].
Furthermore, BI is a more general concept for improving business performance and business decision-making. Big data analytics is a pivotal part for developing BI from a technolo- gical viewpoint and data viewpoint. From a technological viewpoint, big data analytics is big data-driven and business- oriented technology and facilitates business decision-making and then improves BI [6, 7]. From a data viewpoint, big data analytics relies on data analytics and big data which have become a strategic natural resource for every organization, in particular for multinational organizations as well as for e-commerce and e-services. Discovering information, knowl- edge, and wisdom from databases, data warehouses, data marts, and the Web is not only a central topic for business operations, marketing, and BI [7], big data analytics can also use it to measure temporality, expectability, and relativity of intelligence of BI for improving business decision-making.
BASOA: Big data analytics service-oriented architecture
This section proposes a big data analytics service-oriented architecture and then examines each of the main players in the BASOA.
Different from the traditional SOA [34], the proposed BASOA specifies general services to big data analytics services, as shown in Figure 3. We use BA (big analytics) in this architecture, BASOA, to represent big data analytics. This is reasonable because big data and big analytics both are originally from data and analytics, respectively [7, 16].
In this BASOA, big data analytics service provider, big analytics service requestor, and big data analytics service broker are three main players. In what follows, we will look at each of these in some detail, taking BI into account.
In BASOA, big data analytics service requestors include organizations, governments, and all level business decision makers such as CEO, CIO, and CFO as well as managers. Big data analytics service requestors also include business information systems and e-commerce systems. Big data ana- lytics service requestors require big data analytics services including information analytics services, knowledge analytics services, organization analytics services, business analytics services, and market analytics services with visualization tech- niques to provide knowledge patterns and information as well as wisdom [19] for decision-making in the form of figures or tables or reports [35]. More generally, big data analytics service requestors include people who like to make decisions or acquire information based on analytical reports provided by big data analytics service provider [6]. Therefore, a person with smartphone receiving analytics services like GPS infor- mation is also a big data analytics service requestor [7, 17].
Big data analytics service brokers are all the entities that facilitate the development of big data analytics services, which include popular presses, traditional media and social media, consulting companies, scholars and university students, and so on [6]. All these use a variety of methods and techniques to improve the better understanding of big data analytics ser- vices in general and data analytics, business analytics, web analytics, and their services in particular [6]. All these have been offered to university students as a course material or content in business and computing areas to some extent in recent years. McKinsey Consulting (http://www.mckinsey. com/), Boston Consulting Group (BCG), and IDC as big data analytics service brokers have played an important role in pushing big data analytics in businesses and enterprises, just as they promote “big data” [36]. Gartner and Forrester are also famous big data analytics service brokers in the world [17].
Big data analytics service providers include analytics devel- opers, analytics vendors, analytics systems or software, and other intermediaries that can provide analytics services [7]. For example, Tableau as a software developer has been pro- moting big data and big analytics [3]. Recently, web analytics service (WAS) providers are important big data analytics service providers [7]. A WAS provider, for example, Adobe Marketing Cloud (http://www.adobe.com/au/solutions/digi tal-marketing.html), aggregates and analyses blog data about the online behaviors of users who visited the client’s website, then they evaluate a variety of analytical reports concerning the client’s customer online behaviors that the client wishes to understand. This can then facilitate their strategic business decision-making [37]. Application service providers (ASPs) can also provide web analytics services in a hosted ASP model with quicker implementation and lower administrative costs [37]. Analytics developers provide analytic tools with extensive data extraction, analytics, and reporting functional- ity such as Piwik, CrawlTrack [25]. Google is not only a search engine provider, but also a WAS vendor. Google Analytics (http://www.google.com/analytics/) is a big data analytics with good tracking tools. In fact, most hosting web- sites, like Baidu, also provide these similar big data analytics services. A mobile phone company can provide big data analytics services to the customers with smartphone [17]. For example, Mobile App Analytics (http://www.google.com/ analytics/mobile/), a part of Google Analytics, is also a mobile big data analytics services provider that helps the smartphone customers to discover new and relevant users through traffic sources reports. Mobile App Analytics plays a role of integra- tion and gets engaged through event tracking and flow visua- lization, and sets and tracks the goal conversions one wants most: purchases, clicks, or simply time spent on the app [6]. More generally, many information systems have contained an analytics app as a system component to generate tables, dia- grams, or reports. All these kinds of information systems can be considered as big data analytics service providers. Amazon, Google, and Microsoft are examples of big data analytics services providers on the Web [6, 7].
Applying BASOA to enhance BI
This section looks at how to apply the proposed BASOA to enhance BI in some detail.
BAaaS (Big data analytics as a service), as discussed in the BASOA above, means that an individual or organization or information system or software agent uses a wide range of analytic tools or apps wherever they may be located [17]. BAaaS has the ability to turn a general analytic platform into a shared utility for an enterprise or organization with visualized analytic services [17]. A big data analytics service can be available on the Web or used by smartphone [7]. Therefore, big data analytics services include e-analytics ser- vices or WAS and Amazon Web Services (AWS) [3, 6]. Furthermore, big data analytics services also include business analytics services, marketing analytics services, organizational analytics services, security analytics services, and predictive analytics services [5]. Big data analytics services are gaining popularity rapidly in business, e-commerce, e-service, and management in recent years. For example, big data analytics services model has been adopted by many famous web com- panies such as Amazon, Microsoft, and eBay [17]. The key reason behind it is that the traditional hub-and-spoke archi- tectures cannot meet the demands driven by increasingly complex business analytics [17]. BAaaS promises to provide decision makers with visualizing much needed big data [7]. Cloud analytics is an emerging alternative solution for big data analytics [6].
As previously defined, BI is a set of theories, methodolo-
gies, architectures, systems, and technologies that support business decision-making with valuable data, information, and knowledge. BASOA is an architecture for supporting business decision-making with big data analytics services. The theory of big data analytics providers, brokers, and requestors of the BASOA can facilitate the understanding and development of BI and business decision-making. For example, from an in-depth study of the BASOA, an enterprise and its CEO can know who are the best big data analytics providers and brokers in order to improve his organization, business, market performance, and global competitiveness.
We surveyed 71 information technology managers at the Association for Education in Journalism and Mass Communication (AEJMC) in Montreal during August 6–9, 2014 [6], to collect data concerning the enterprise-level acceptability of the BASOA concept. These results indicate some preliminary support for the BASOA concept of having service brokers work with service requesters and providers similar to the way private mortgage and loans work in the USA. Based on this preliminary enterprise acceptability of this BASOA model, we propose that more research be done to investigate how it could be used in the near future.
Related work and discussion
We have mentioned a number of scholarly researches on data analytics, big data analytics, and BI. In what follows, we focus on related work and discussion on ontology of big data analytics, and the work of SAP as well as incorporation of big data analytics into BI.
Why does big data analytics really matter for modern business organizations? There are many different answers to this question from different researchers. For example, Davis considers that the current big data analytics has embodied the state-of-the-art development of modern computing [38], which has been reflected in “An ontology of big data analy- tics.” Gandomi and Harder [14] discuss how big data analytics has captured the attention of business and government leaders through decomposing big data analytics into text analytics, audio analytics, video analytics, social media analytics, and predictive analytics. This implies that big data can be classified into big text data, big audio data, big video data, and big social media data [14].
Big data analytics and BI have drawn an increasing atten- tion in the computing, business, and e-commerce community recently. For example, Lim et al. [11] examine BI and analytics by focusing on the research directions. They consider BI and analytics (BIA) as a current form replacing the traditional BI, whereas we still consider BI and big data analytics are two different concepts, although they have close relationships and share some commons. One of the contributions of this article is that big data analytics services can enhance BI. Fan et al.
[14] provide a marketing mix framework for big data manage- ment through identifying the big data sources, methods, and applications for each of the marketing mix, consisting of people, product, place, price, and promotion. However, they have not shown the relationship between marketing intelli- gence and BI in terms of big data analytics.
Ontology has been important in computer science and AI [22]. A basic search in Google scholar (i.e. article title and key words) reveals that there are few publications entitled “ontol- ogy of big data analytics.” We then explored it and put it as a part of this research through updating our early work on data analytics, business analytics, and big data analytics [6, 7]. Comparing with the early proposed ontology of big data analytics [7], the proposed ontology of big data analytics in this article, illustrated in Figure 1, includes a technology level consisting of DW + DM + SM + ML+ Visualization + Optimization. All these reflect the latest progress of ICT in general and in DW, DM, SM, ML, Visualization, and Optimization in specific. Incorporating these advanced tech- niques into data analytics for big data is the essence of big data analytics [7]. More specifically, this added level arms big data descriptive, predictive, and prescriptive analytics with these advanced techniques in ICT. This is only a beginning for providing a relatively comprehensive ontology of big data analytics. In this direction, we will investigate more academic reviewed sources as a future work to develop an ontology of big data analytics with three levels for each related analytics: big data, methods, and applications based on the method of Fan et al. [14]. Such an investigation would become an important guide for the research and development of big data analytics.
The introduction of temporality, expectability, and relativ-
ity as three dimensions of BI is a new understanding of intelligence in BI. This is significant to decompose intelligence in BI into the smaller ones, just as we decompose materials to molecules to atoms to protons, etc. Further, we can narrow our focus on these three dimensions of BI. However, how big
data descriptive analytics, predictive analytics, and prescrip- tive analytics [7, 16] are all connected to temporality, expect- ability, and relativity of intelligence in BI is a big issue. The big data analytics services can improve the temporality, expectability, and relativity of intelligence in terms of systems, products, and services as well as organizations and then improve the decision-making of managers so that the BI has been enhanced. This will be extended in the future research paper entitled “Temporality, expectability and relativity of business intelligence for big data analytics”.
SAP, one of the leading vendors of ERP [39], has intro- duced its enterprise service-oriented architecture (ESOA) [25]. SAP’s ESOA specifies general services to enterprise ser- vices whereas our BASOA model specifies general services to big data analytics services. Big data analytics services should be a part of state-of-the-art e-commerce, e-services, and the Internet of services (IoS) [23], and then the proposed BASOA can be considered as a concrete application for the ESOA of SAP. However, SAP’s enterprise systems focus on key applica- tions in finance, logistics, procurement, and human resources management as an ERP system. We conceive that our BASOA will be incorporated into the next generation enterprise sys- tems integrating SCM, CRM, and KM systems, and e-com- merce systems. This is also the motivation of our proposed BASOA.
Conclusion
This article examined how to use big data analytics services to enhance BI by presenting an ontology of big data analytics and a BASOA, and then applying BASOA to BI, where our surveyed data analysis showed that the proposed BASOA is viable for enhancing BI and enterprise information systems. This article also examined temporality, expectability, and relativity as the characteristics of intelligence in BI, and dis- cussed the interrelationship between BI and big data analytics. The proposed approach in this article might facilitate the research and development of business analytics, big data ana- lytics, BI, e-services, and IoS as well as big data computing and big data science.
In the future work, besides mentioned in the previous sections, we will analyze the foregoing collected data vigor- ously and explore enterprise and e-commerce acceptability of BASOA for BI. We will also realize BASOA using intelligent agents’ technology [40], where we will look at some imple- mentation-related issues such as how to collect, store, and process big data—by whom, for what, access rights, and many more. We will provide significant examples for model- ing temporality, expectability, and relativity of intelligence in BI using big data analytics such as Google Analytics.
References
[1] Reddy CK. A survey of platforms for big data analytics. J Big Data (Springer). 2014;1(8):1–20.
[2] Chen CP, Zhang C-Y. Data-intensive applications, challenges, techni- ques and technologies: a survey on Big Data. Inf Sci. 2014;275:314–347.
Tableau. 2015. Top 8 Trends for 2016: Big Data. [Online]. Available from www.tableau.com/Big-Data.
[1] McAfee A, Brynjolfsson E. Big data: the management revolution. Harv Bus Rev. 2012;90(10):61–68.
[2] Roche S. IDC reveals 53% of organizations in the APEJ region con- sider big data and analytics important for business. 21 Apr 2016. [Online]. Available from http://www.idc.com/getdoc.jsp? containerId=prAP41208316
[3] Sun Z, Strang K, Yearwood J. Analytics service oriented architecture for enterprise information systems, in Proceedings of iiWAS2014, CONFENIS 2014, 4–6 Dec 14, Hanoi, 2014.
[4] Sun Z, Zou H, Strang K. Big data analytics as a service for business intelligence. in I3E2015, LNCS 9373, Berlin, 2015.
[5] Sun Z, Firmin S, Yearwood J. Integrating online social networking with e-commerce based on CBR. in The 23rd ACIS 2012 Proceedings, 3–5 Dec, Geelong, 2012.
[6] Ali Z. New IDC MarketScape provides a vendor assessment of the worldwide business analytics consulting and systems integration services for 2016. 29 Apr 2016. [Online]. Available from http:// www.idc.com/getdoc.jsp?containerId=prUS41224416.
[7] Vesset D, McDonough B, Schubmehl D, Wardley M. Worldwide Business Analytics Software 2013–2017 Forecast and 2012 Vendor Shares (Doc # 241689). 6 2013. [Online] [cited 2014 Jun 28]. Available from http://www.idc.com/getdoc.jsp?containerId= 241689
[8] Lim E, Chen H, Chen G. Business intelligence and analytics: research directions. ACM Trans Manage Inf Syst. 2013;3(4):1–10.
[9] Luhn HP. A business intelligence system. IBM J Res Dev. 1958;2 (4):314–319.
[10] Turban E, Volonino L. Information technology for management: improving strategic and operational performance (8th Ed). Danvers, MA: Wiley & Sons; 2011.
[11] Fan S, Lau RY, Zhao JL. Demystifying big data analytics for business intelligence through the lens of marketing mix. Big Data Res. 2015;2:28–32.
[12] Gandomi A, Haider M. Beyond the hype: big data concepts, methods, and analytics. Int J Inf Manage. 2015;35:137–144.
[13] Minelli M, Chambers M, Dhiraj A. Big data, big analytics: emer- ging business intelligence and analytic trends for today’s busi- nesses. Wiley & Sons (Chinese Edition 2014, Posts & Telecom Press); 2013.
[14] Delena D, Demirkanb H. Data, information and analytics as services. Decis Support Syst. 2013;55(1):359–363.
[15] Kantardzic M. Data mining: concepts, models, methods, and algorithms. Hoboken, NJ: Wiley & IEEE Press; 2011.
[16] Coronel C, Morris S. Database systems: design, implementation, and management (11th Ed). Boston, MA: Cengage Learning; 2015.
[17] Cramer C. How descriptive analytics are changing marketing. 19 May 2014. [Online] [cited 2015 July 6]. Available from http:// www.miprofs.com/wp/descriptive-analytics-changing-marketing/
[18] Gruber T. Toward principles for the design of ontologies used for knowledge sharing. Int J Human–Computer Stud. 1995;43(5– 6):907–928.
[19] Sun Z, Yearwood J. A theoretical foundation of demand-driven web services. In Sun Z, Yearwood J, editors. Demand-driven web services: theory, technologies, and applications. IGI-Global; 2014, p. 1–25.
[20] Sabherwal R, Becerra-Fernandez I. Business intelligence: practices, technologies, and management. Hoboken, NJ: John Wiley & Sons Inc.; 2011.
[21] Holsapple C, Lee-Post A, Pakath R. A unified foundation for business analytics. Decis Support Syst. 2014;64:130–141.
[22] Laudon K, Laudon J. Management information systems: mana- ging the digital firm (14th Ed). Boston, MA: Pearson; 2016.
[23] Wang F-Y. A big-data perspective on AI: Newton, Merton, and analytics intelligence. IEEE Intell Syst. 2012 Sept/Oct;27(5):2–4.
[24] Russell S, Norvig P. Artificial intelligence: a modern approach (3rd Ed). Upper Saddle River, NJ: Prentice Hall; 2010.
[25] Macmillan. Macmillan English dictionary for advanced learners. London: Macmillan; 2007.
[26] Negnevitsky M. Artificial intelligence: a guide to intelligent sys- tems (2nd Ed). Harlow: Addison-Wesley; 2005.
[27] Astrom KJ, McAvoy TJ. Intelligent control. J Process Control. 1992;2(3):115–126.
[28] Larson EK, Gray CF. Project management: the managerial process (5th Ed). New York: McGraw-Hill; 2011.
[29] IDC. IDC Predictions 2014: Battles for dominance — and survival
— on the 3rd Platform. Dec 2013. [Online] [cited 2014 Feb 13]. Available from http://www.idc.com/getdoc.jsp?containerId= 244606
[30] Brust A. Gartner releases 2013 BI Magic Quadrant. 2013. [Online] [cited 2014 Feb 14]. Available from http://www.zdnet.com/gart ner-releases-2013-bi-magic-quadrant-7000011264/
[31] Papazoglou MP. Web services: principles and technology. Harlow, England: Pearson Prentice Hall; 2008.
[32] Kauffman RJ, Srivastava J, Vayghan J. Business and data analy- tics: new innovations for the management of e-commerce. Electron Commerce Res Appl. 2012;11:85–88.
[33] McKinsey. Big data: the next frontier for innovation, competition, and productivity. May 2011. [Online]. Available from http://www. mckinsey.com/business-functions/business-technology/our- insights/big-data-the-next-frontier-for-innovation
[34] Park J, Kim J, Koh J. Determinants of continuous usage intention in web analytics services. Electron Commerce Res Appl. 2010;9(1):61–72.
[35] Davis CK. Viewpoint beyond data and analytics: why business analytics and big data really matter for modern business organiza- tions. CACM 2014;57(8):39–41.
[36] Elragal A. ERP and big data: the inept couple. Procedia Technol. 2014;16:242–49.
[37] Sun Z, Finnie G. Intelligent techniques in e-commerce: a case- based reasoning perspective. Heidelberg, Berlin: Springer-Verlag, 2004; 2010.
[38] van der Meulen R, Rivera J. Gartner says Worldwide Business Intelligence and Analytics Software market grew 8 percent in 2013. 29 Apr 2014. [Online] [cited 2014 Jun 28]. Available from http://www. gartner.com/newsroom/id/2723717
[39] Schalkoff RJ. Intelligent systems: principles, paradigms, and prag- matics. Boston, MA: Jones and Bartlett Publishers; 2011.
[40] Sun Z, Strang K, Firmin S. Business analytics-based enterprise infor- mation systems. J Comput Inf Syst. 2016. doi:10.1080/ 08874417.2016.1183977.
