Role of the Big Data Analytic framework in Business Intelligence and its Impact: Need
and Benefits
Farhad Khoshbakhta, Atena Shiranzaeib, S. M. K. Quadric
a Department of Computer Science Jamia Millia Islamia (A Central University), New Delhi, India 110025 bDepartment of Computer Engineering, Faculty of Industry and Mining(Khash),University of Sistan and
Baluchestan,Zahedan,Iran
cDepartment of Computer Science JamiaMilliaIslamia(A Central University), New Delhi, India 110025 a f.khoshbakht630@gmail.com, b ashiranzaei@eng.usb.ac.ir, c quadrismk@jmi.ac.in
Article History: Received: 10 January 2021; Revised: 12 February 2021; Accepted: 27 March 2021; Published
online: 28 April 2021
_____________________________________________________________________________________________________ Abstract: Big Data is mixed with huge, autonomous sources like decentralized and distributed control system. For
organizations that use the conventional data processing mechanism to handle and archive these large data sets, these capabilities pose an extreme obstacle. A new model has to be defined and the existing framework must be re-evaluated for the analysis and management of big data. The word Business Intelligence (BI) applies to applications, technology and activities for commercial knowledge gathering, review, incorporation and presentation. Business Intelligence’s primary aim is to facilitate quicker and stronger business decision-making process.” Therefore, the strategic review of the literature must study the trend of Information System (IS) adoption factors as a well-designed strategic diagnostic tool that can be used for essential decision-making systems to enable more effective and reliable action plans. We started with discussing some frameworks required for strategic excellence by examining the potential approaches of Big Data Analytics (BDA) and Business Intelligence (BI). In the end, we would design an integrated application that functions as an organization's strategic performance management diagnostic tool. “In general, the emphasis and reach are on the corporate decision-making mechanism, there are some specifics on the analysis, but every conceivable tool and instrument is not specified, the concept is sufficiently concrete to assist in the creation of steps. This research paper therefore examines the position of the system for big data analytics and market intelligence.
Keywords: Big Data, Business Intelligence, Big Data Analytics
1. Introduction
Today web mining is a requesting task in the organization. Every organization created a lot of data from different sources. Web mining is the way toward removing helpful information from web assets. Log documents are overseen by the Web server. The difficult errand for Business Intelligence is to know client conduct to improve business by breaking down Web log records. The examination of Big Data has been received as a problematic technology that will change the business intelligence, which is a domain that depends on data investigation for business bits of knowledge. for better decision making.
I).Big Data
The “Big data relates to data sets whose dimensions go outside the scope of the running of the programming machinery of the mill database to collect , store, track and break down knowledge (Uma Narayanan et al., 2020). Petabytes (1 petabyte = 1024 terabytes), exabytes (1 exabyte = 1024 petabytes), zettabytes (1 zettabyte = 1024 exabytes), yottabytes (1 exabyte = 1024 yottabytes) (Maryam AitHadj et al., 2019) are increasing as data sizes increase.”
Big Data's Features
Big data may be represented from "a few points of view." The angles commonly used are volume, velocity and variety. In comparison, Reality and Value are used to describe Big Data. They are beneficial aims from which we can grasp the Big Data definition and the open stage for their use.
Volume: as the system turns out to be increasingly open and available, the data provided from different sources is exceptionally huge; zettabyte or petabyte. This massive data calculation is called Big Data.
Velocity: The basic engine of Big Data is the supreme level at which we generate data. Data produced from numerous sources runs from batch to real time. So this quick data characterises another concept named Big Data.
Figure 1 Big Data's Features
Variety: the representation of data provided by multiple sources is of an alternative type; for example, organised data is supervised by online business pages, semi-organized data is supervised by web server logs and unstructured data is supervised by social sites, for example, video, sound, pictures, and so forth face everybody's blend.
Veracity: I pair at the total speed of certain data where we can't sit around cleaning them before utilizing them. Compiling data from numerous sources and utilizing them for business decision making requires a component that manages off base data. The combination of accurate, off-base, and accurate data will also be considered big data.
Value: by handling a lot of volume, fast, honesty and variety of data, it presents another measurement for the investigation of big data called "value". Teaming up numerous kinds of data, uniting them all to separate concealed business bits of knowledge and increase an upper hand from them speaks to the value of big data. II). Big Data Analytics
Data analysis coordinates "the information received on the occasion of intrigue through perception, measurement or trials." The purpose of data analysis is to separate as much information as could reasonably be expected to be relevant to the viable subject matter. In addition , the entire field of big data analysis was ordered (Mishra et al., 2018) into three levels depending on the depth of the analysis: descriptive analysis, predictive analysis and prescriptive analysis, description as given below:”
A. Descriptive analysis:
To represent what happened, it uses recorded evidence. For e.g., a relapse may be used to find simple trends in the databases, the representation simply introduces the data, and the data demonstration is used to capture, store and cut data skillfully.
B. Predictive analysis:
It centers on foreseeing future patterns and probabilities. For instance, predictive displaying utilizes factual techniques, for example, straight relapse and coordination’s to get inclines and anticipate future outcomes, while data extraction separates models to give information and expectations.
C. Prescriptive analysis:
This discusses decision making and integrity. For example, reproduction is used to analyze complex systems to gain details on system actions and to discriminate between problems, and optimization techniques are utilized to discover ideal arrangements dependent on specific requirements.”
III). Business Analytics
Market Analytics is a systemic way of thought in which subjective, quantitative and mathematical computing devices and strategies are used to break down knowledge, collect information, illuminate and improve decision-making. A number of methods, including indicative, predictive, and prescriptive and optimization frameworks, may be used in a complex study. Relevant research subtypes cover both market analysis and data analysis. The
principles of descriptive, predictive and prescriptive analysis allude to the form of model and each of the three subtypes of market and data analysis are shared. Most issues happen in IS decision making for key decision making from different points of view, to progress past restricted organizational ancient rarity contemplations and to attempt ineffectively investigated BDA organizational settings. This prompted genuine performances and loss of intensity.
Numerous ventures, for example, account, internet based life, social insurance, security, record mining, are grasping big data technologies with the guarantee of extending the capacity to handily extricate a lot of data. As future work, it will be fascinating to design the big data analysis motor in real time and perceive how the data oversee Hadoop MapReduce, contrast its performance and the appropriated group preparing engineering and see how it conquers difficulties in the system Batch handling of Big Data Analytics.
IV). “Business Intelligence”
Business Intelligence ( BI) is a non-exclusive concept that encompasses technologies, foundations and devices and best practises that enable knowledge to be obtained and evaluated to strengthen and optimise decisions and execution, according to Gartner (EGHAM, U.K., 2019). It is a form of programming that is intended to report, break down and talk to data intermittently. A large portion of the data is perused from the data stockroom or data bazaar. Enter data mining based on knowledge and data mining based on technologies and solve any problems between market intelligence information and the numerous data mining techniques in web-based business.” 2. Objectives
• To differentiate between frameworks of Big Data Analytics (BDA) and Business Intelligence (BI) approaches
• To explore the Real Time Big Data Analytics and its impact on business performances. • To study the need and benefits of big data analytics in business intelligence (BI).
3. Vision
Many organizations are so interested in striving for excellence. However, these efforts are not easy. Some are still interacting with these operational difficulties and putting strategic scenarios at risk. Many have experienced obstacles and risks arising from huge data silos - isolated information repositories. Hence the vision of this article to emphasize the role of the Big Data framework for the analysis of business intelligence and its impact.
4. Role Of Big Data Analytics Frameworks
Various forms of knowledge utilised when considering Big Data. To conduct different types of research, different types of frameworks are needed. A number of workloads in the field of large-scale data processing. We also see a mixture of these dispersed workloads in order to accomplish a company purpose:
• For instance, Map Reduce dependent frameworks such as Hadoop, for recurring tasks such as large-scale data mining or aggregation, the' Batch-oriented processing
• With Apache HBase Online Transactional Processing (OLTP), such as user-facing e-commerce purchases,
• Stream processing, with Storm being a representative system for managing stream sources such as social network feeds or sensor info,
• Interactive ad-hoc query with Apache Drill and review.” 1). “Apache Hadoop Method”
Apache Hadoop is an open source software library that implements a structure that takes into account the distributed preparation of immense data sets using simple programming models in PC collections. You have a range of options, from a single PC to a vast number of PCs, each providing handling and capability in the vicinity. The library itself is intended to recognise and track disappointments and maintain high usability at the application stage, rather than relying on facilities.”
The “Apache Hadoop include following modules:
a) Hadoop core: basic utilities supporting additional modules
b) Hadoop Distributed File System: Allows access to programme data with fast throughput. c) Hadoop Map Reduce: Broad Data Collection Parallel Processing System.”
d) Hadoop YARN: Task Preparation and Inventory Control system 5. Literature Review
Business analytics (BA) applies to expertise, technologies, rehearsals for continuous iterative investigation and the search for knowledge and direct business structure for previous corporate success (Vincent Whitelock 2018). Organizations concerned with data-driven decision-making utilizingbusiness intelligence. BA is used to collect knowledge that informs organization decisions that can be used to robotize that develop company types. Data-driven businesses regard the knowledge as a business commodity and influence it to improve an upper hand. Fruitful market research depends on trained examiners who appreciate data-driven decision-making technologies and organization and corporate accountability. Instances utilized by BA include: research discovery to find new models and similarities (data mining), clarification of whether a particular outcome resulted (statistical analysis, quantitative analysis), potential outcome predicting (predictive analysis).”
As indicated by Phillips-wren, (2015), Business Intelligence is introduced as an application stage to help business decisions by featuring the systematic procedure for unstructured data, data sources and buildings. the workshop of the specific vested party on decision emotionally supportive networks (SIGDSS), individually. Perceiving the capability of "big data" to offer new thoughts for decision making and development, the speakers at the two occasions talked about how organizations can utilize and oversee big data to increase an upper hand. Then again, Business Intelligence and Big Data Analytics have risen as diagnostic, specialized, building and application devices to aid key decision making (Jung and Wu, 2016). Big Data Analytics is predictive, while Business Intelligence helps in educated decision making dependent on the analysis of past data. The point of the examination is to decide the reception of the hypothetical framework towards the applied framework utilizing the BI capacity to break down the nature of the data introduced as KPI from operational administration to BDA.
Business analysis utilizes data, statistical and quantitative examines and informative and predictive models to help settle on noteworthy decisions and improve business activities. There are numerous kinds of business analysis, constant versus non-genuine, key versus strategies, arranged versus spontaneous and organized versus unstructured (Chan, 2012).
Chiefs “have utilized business analysis to advise their decision making for quite a long time. They are currently utilizing business analysis not exclusively to examine past performance yet in addition to distinguish chances to improve future performance and its challenges. As indicated by UthayasankarSivarajah et al. 2017 business analysis comprises of big data analysis, test analysis, web analysis, organize analysis and mobile analysis, huge numbers of which are unstructured and can't be broke down utilizing the executives devices. a few data social databases.”
Numerous organizations are so inspired by the quest for greatness. Many have encountered snags and dangers emerging from colossal data storehouses - detached information vaults. “IS has become most organizations are founded as an interconnected and facilitated set of components that join together to turn data into knowledge (Jaques, 2017). It is identified as the programme that analyses and extracts knowledge (Rahman et al., 2017). The main goal of IS is to turn evidence into material knowledge that can be used in an organization to make decisions (Murugesan and Karthikeyan, 2016).”
Vital decision making is a nonstop procedure of making the organizations crucial, goals, targets and an essential segment of the organization's administration for a specific activity of the system alteration plan dependent on the outcomes watched (Kohtamäki and Farmer, 2017). The vital decision gives a basic appraisal of the connection between decision making and OL performance (Saadat and Saadat, 2016). Moreover, these organizations are encountering redundancies and data blunders, just as a bottleneck and information over-burden.
A few specialists distinguished these inquiries, analyzed “writing and characterizing BA (once in a while in conjunction with big data or market intelligence) as a creation in the field of IS, to be viewed using the frameworks of that sector. Grover, Chiang, Liang and Zhang (2018) addressed "big data and research" on how corporate value is generated. A structure for the 'knowledge value chain' was suggested by Abbasi, Sarker and Chiang (2016).look into with regards to big data, utilizing three IS examine customs ("IS behavior, design and financial aspects"). ShahriarAkter et al. (2019) anticipated a pattern from diagnostically of analytic based decision making for making service system and guide for industries practitioners.”
Mobile telephones have become a successful station for arriving at “Various consumers and as a means to increase the effectiveness and competitiveness of the employees of a company. In addition, the Hype Period review of Gartner BI found Smartphone BI as one of the emerging developments that might probably upset the BI showcase (EGHAM, U.K., (October 2, 2019). With the accelerated growth of mobile devices and increasingly virtual terminals, for instance, all inclusive tablets, cameras, RFID and applications are being executed. Late
advancements in remote controls, mobile devices and transmission handling have spurred the introduction of body sensor systems to continuously track the well-being of an individual.”
Through incorporating a big data analysis system that speaks to a process perspective on the segments needed for big data analysis in organizations, we transcend every obstacle between scholarly and proficient science. Both equal handling and the method of carrying calculation to data made it conceivable to process enormous data sets at rapid. These key highlights and the capacity to process a lot of data were an incredible inspiration for breaking down the design of Apache's industry-driving data preparing framework, Hadoop.” See how this tremendous measure of data and analysis is cultivated, and experience the RDBMS versus condition. Hadoop has been appeared to give an inside and out perspective on the most talked about technology. We discern the ebb and flow circumstance of the discovery of big data guided by the system and suggest possible territories for future study to establish the value of scholastic study by and by.
6. Need And Benefits Of Big Data Analytics In Business Intelligence (Bi)
Data is monstrous, arrives at a pace and fundamentally unstructured that does not match conventional social database systems, "the big data features alluded to above." An elective method of preparing this monumental data is necessary with too much knowledge covered in this data. Huge organizations may have the assets to deal with this undertaking, yet the measure of data that is created each day effectively surpasses this limit. Cheaper computers, cloud storage, and open source software also allowed a far lower cost of managing big data. Bunches of data imply tonnes of shrouded data. The ability to break down big data quickly means the ability to find out about consumers, spotlight trends, marketing and promotional efforts, community observation and results review, and dramatically more. Furthermore, this is a significant motivation behind why numerous huge organizations need strong big data analytics apparatuses and technologies.”
Big Data devices basically utilize the standard of questioning data in memory. In comparison to traditional business a intelligence (BI) application that runs problems against details stored on the hard drive of the computer, queries are rendered where the knowledge is stored away. In-memory data processing also increased the accuracy of data problems altogether. Big Data analytics not only enable companies create smarter choices and add a little leeway to manage them increasingly, however has additionally enlivened organizations to infer new measurements and determine new wellsprings of income from bits of knowledge picked up. Remember that transient data normally prompts Big Data, much the same as spatial data. Low cost and extremely reliable strategies for big data for spatial temporal interpolation [Esmaeilbeigia et al., 2020], i.e. entity relationship databases, is used to handle supermarket chains, including non-scalar data. Big Data beats ORDBMS in a few different ways, including the requirement for increasingly entangled backups, recovery, and quicker hunt calculations, past RDBMS files. The advantages of utilizing Big Data Technologies can have the hindrance of lost data protection. As far as security, a few organizations offer client data to different organizations, and this can be an issue. A lot of unstructured data is produced by company and logical analysis fields. The control of this structured knowledge depends on the right RDBMS, data storage, BPM and OLAP. In essence, data processing is focused on data mining and mathematical analysis. Process mining has been another form of concentrated modification and study. Despite market frameworks produced for RDBMS, segment-based DBMS, in-memory DBMS, and equal DBMS (Georgios 2017), Hadoop and MapReduce-based frameworks has been another rational option for Big Data Analysis.”
7. Role Of Real Time Big Data Analytics And Its Impact
Only a couple of years prior, creating a question result on petabytes of data in under an hour were viewed as a marvel. Be that as it may, innovative advances have permitted us to get brings about not exactly a moment. Consider a question, get an outcome, and start the trial.
Albeit big data analytics help settle on data-driven decisions, big data analytics applications are presently connected by huge inertness saw by end clients. Practically all generally utilized “Big Data technologies are not reasonable for real-time analysis. The assignment is as troublesome as finding the needle in the bundle in a matter of seconds.” The issue is additionally disturbed when the data can be connected to other data [Ashwini et al., 2017].
So far a large portion of “the consideration on Big Data has concentrated on circumstances where the data to be questioned has just been gathered and put away in a Big Data database. The analysis of Big Data in real time, then again, attempts to examine data in constant advancement: the transmission of data.” Occasions/information are ceaselessly sent to the framework and dissected based on bits of knowledge taken from the data previously gathered. It would not be sensible to chronicle all occasions and anticipate that the RTBDA framework should give answers in milliseconds. Along these lines, the RBDTA framework works by examining occasions without losing the data value. How quick is it sufficiently quick? “Quick is a relative term, various frameworks/situations
may have various desires. Getting a traffic update two minutes after the fact may not be as awful as beginning an activity after 10 milliseconds.”
The “Real Time Big Data Analytics discovers application in a heap of divisions, including normal speculates, for example, Algorithmic Trading and Healthcare. How about we inspect a portion of the applications top to bottom.
• Financial industry: the budgetary business is one of the most severe ventures where time is on the rundown of the most persuasive components. Standing by excessively long and anticipating that a contender should pass up on the chance meanwhile may not be suggested. Exchanging robots (calculations) must concentrate gigabytes of data and must enact (or decide not to actuate) a trade in milliseconds dependent on current economic situations. Dealers likewise need to separate a great deal of data to show signs of improvement goads in the securities exchange (Ketaki 2016).
• Fraud detection: RTBDA can be utilized to identify fraudulent exchanges. In the event that the fraud detection framework recognizes that the client has made an exchange some place on the east coast and inside 5 minutes there is an exchange on the east coast, an alert is activated. This would spare a huge number of dollars every year (Ketaki 2016).
• Traffic refreshes and steering: envision a world wherein you won't be stuck in blocked rush hour gridlock, wondering why you chose to take the way wherein you stuck, depending on cell phones for route applications that nearly everybody and the capacity to rapidly misuse the huge transmission of data produced by these route applications. The greater part of these applications works admirably of giving various courses and permit the client to choose a course. In any case, traffic circumstances change ceaselessly and the greater part of the present route frameworks don't consider traffic blockage/mishaps to furnish the client with a superior elective course (Ketaki 2016).
• Framework Monitoring/Log Mining - Current applications/bunches are getting progressively mind boggling, making it hard to screen and trigger cautions if the mistake rate and/or framework performance contrasts. Log mining frameworks like Splunk, New Relic can get log streams from a huge number of hubs running applications in real time. The real-time feed is contrasted with authentic data to search for any deviations from typical behavior. Any deviation over a set limit will be viewed as a potential issue and the alert will go off (Ketaki 2016).”
8. Conclusion
With the approach of computerized technology and brilliant gadgets, “every day, a lot of advanced data is generated. To this immense measure of details, developments in computerized sensors and correspondence technologies have contributed enormously, gathering valuable information for corporations and companies. Using conventional systems, these big data are challenging to process and need tremendous egalitarian planning. Technologies that can store and procedure exabytes, terabytes, petabytes of data without fundamentally expanding the cost of data storage are a need of time. The capacity to acquire information from these big data can possibly change the manner in which we live, think and work. The advantages of big data analysis extend from the domain of wellbeing to government, fund, advertising and the sky is the limit from there. Open source technologies for Big Data have accomplished significant footing because of the demonstrated capacity to process huge amounts of data in equal.
The “Hadoop is the most generally acknowledged and utilized open source framework for computing big data analysis in an effectively adaptable condition. It is a modest, solid, exceptionally adaptable and reasonable arrangement that bolsters the spread of equal bunches on a huge number of hubs and can oversee petabytes of data. Two fundamental parts HDFS and MapReduce add to Hadoop's prosperity. It handles the storage and analysis of unstructured data well indeed. Hadoop is a demonstrated arrangement in the creation condition and very much received by significant industry organizations, for example, Google, Yahoo and Facebook. Albeit past renditions of Hadoop didn't make some real-memories data analysis part, Apache as of late presented Spark as an answer for breaking down real-time big data. Spark depends on strong conveyed data and is said to convey brings about a brief instant.”
The study recommends the organization to provide employees with Big Data training during the training strategy. This will allow employees to update themselves with information about the data on which decisions can be easily made through data interpretation. The organization must have adequate structures for Big Data analysis structures to improve efficiency, promote teamwork, create synergies, and reduce costs in support of its final strategy in which the organizational structure that develops is based on the organization strategy.
References
1. Abbasi, A., Sarker, S., & Chiang, R. H. L. (2016). Big data research in information systems: Toward an inclusive research agenda. Journal of the Association for Information Systems, 17(2). Article 3. Retrieved from http://aisel.aisnet.org/jais/vol17/iss2/3
2. Ashwini D. Meshram, Anuja S. Kulkarni, and Saanavi S. Hippargi (2017) Big Dara Analytics using Real-Time Architecture. International Journal of Latest Trends in Engineering and Technology, 6(4), 241-247.
3. EGHAM, U.K., (October 2, 2019) Gartner Reveals Five Major Trends Shaping the Evolution of
Analytics and Business Intelligence. Newsroom, Press Releases,
https://www.gartner.com/en/newsroom/press-releases/2019-10-02-gartner-reveals-five-major-trends-shaping-the-evoluti.
4. Esmaeilbeigia M., O. Chatrabgoun, A.Hosseinian-Far, R.Montasari, A.Daneshkhah (July 2020). A low cost and highly accurate technique for big data spatial-temporal interpolation, Applied Numerical Mathematics. 153, 492-502.
5. Georgios K. Gaitanis (2017) Big Data. Thesis, National AndKapodistrian University Of Athens Department Of Informatics And Telecommunications Department Of Economics.
6. Gloria Phillips-Wren, Lakshmi Iyer, Uday Kulkarni, T. Ariyachandra (2015). Business Analytics in the Context of Big Data: A Roadmap for Research, Communications of the Association for Information Systems 37:448-472 DOI: 10.17705/1CAIS.03723
7. Grover, V., Chiang, R. H. L., Liang, T. P., & Zhang, D. (2018). Creating strategic business value from big data analytics: A research framework. Journal of Management Information Systems, 35(2), 388–423. 8. Jaques, E. (2017). Requisite organization: A totalsystem for effective managerial organization and
mana-gerial leadership for the 21st century. Routledge
9. KetakiSubhashRaste (2016) Big Data Analytics – Hadoop Performance Analysis, Master of Science Thesis, Faculty of San Diego State University, p. 40.
10. Kohtamäki, M., Farmer, D. (2017). Real-time Strat-egy and Business Intelligence. p. 11–37
11. MaryemAitHadj, Mohammed Erradi El, Ahmed Khoumsi, YahyaBenkaouz (2019) Validation and correction of large security policies: a clustering and access log based approach Proceedings - 2018 IEEE International Conference on Big Data, Big Data 2018 (2019), pp. 5330-5332,
12. Mishra, P., Pandey, C. M., Singh, U., & Gupta, A. (2018). Scales of measurement and presentation of statistical data. Annals of cardiac anaesthesia, 21(4), 419–422. https://doi.org/10.4103/aca.ACA_131_18. 13. Murugesan, M. and Karthikeyan, K. (2016). Busi-ness Intelligence Market Trends and Growth in
Enterprise Business, International Journal on Recent andInnovation Trends in Computing and Communication, 4(3) 188–192
14. Rahman, M. N. A., Zamri, S. N. A. S., Leong, K. E.(2017). A Meta-Analysis Study of Satisfaction and Con-tinuance Intention to Use Educational Technology, Inter-national Journal of Academic Research in Business andSocial Sciences, 7 (4) 1059–1072.
15. Saadat, V., Saadat, Z. (2016). Organizational Learn-ing as a Key Role of Organizational Success, Procedia -Social and Behavioral Sciences. 230 (May), p. 219–225.
16. ShahriarAkter, RuwanBandara, Umme Hani, Samuel FossoWambac, Cyril Foropond, ThanosPapadopoulose (October 2019). Analytics-based decision-making for service systems: A qualitative study and agenda for future research. International Journal of Information Management, 48, 85-95.
17. Uma Narayanan, Varghese Paul, ShelbiJosepha (2020) A novel system architecture for secure authentication and data sharing in cloud enabled Big Data Environment. Journal of King Saud University - Computer and Information Sciences.
18. UthayasankarSivarajah, Muhammad MustafaKamal, ZahirIrani, VishanthWeerakkody (January 2017). Critical analysis of Big Data challenges and analytical methods.Journal of Business Research, 70, 263-2868
19. Vincent Whitelock (2018). Business analytics and firm performance: role of structured financial statement data, Journal of Business Analytics, 1(2).
