Research and Reviews - International Journals

International Journal of Innovative Research in Computer and Communication Engineering

Menu

ISSN ONLINE(2320-9801) PRINT (2320-9798)

All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.

BUSINESS INTELLIGENCE IN A NUTSHELL

Navita Kumari
  1. Research Scholar, Guru Kashi University, Bathinda, India
Related article at Pubmed, Scholar Google

Visit for more related articles at International Journal of Innovative Research in Computer and Communication Engineering

International Journal of Innovative Research in Computer and Communication Engineering

Abstract

Business Intelligence is gaining popularity in organizations. Business Intelligence systems are developed to help the organizations understand their customers, operations and performance. BI systems act as measurements units for a business or an organization. Poor intelligence leads to poor decision making. BI has problems like data quality, which an organization has to overcome, in order to improve its decision making. A serious effort has been made to understand the cause of poor quality business analytics over the past several years. Most organizations and analysts agree that the main reason of erroneous reporting is the operational data which is used for generating these analytical reports as the data is filled with errors, duplications and inconsistencies. This research paper explores the concepts of BI, data quality characteristics and data quality issues, types of BI tools and delivery mechanisms and the key features of BI architecture.

 

Keywords
Business Intelligence (BI), Data Quality (DQ), Decision Support System (DSS), Enterprise Data Warehouse(EDW), Extract Transform Load (ETL), Key Performance Indicator (KPI), Online Analytical Processing (OLAP).

INTRODUCTION
Business intelligence (BI) is defined as the ability for an organization to take all its processes and capabilities and thenconvert these into knowledge, ultimately getting right information for the right people, at the right time, through the rightchannel. This produces large amounts of information which can lead to the development of new opportunities for theorganization. When these opportunities have been identified and a strategy has been effectively implemented, BI canprovide an organization with a competitive advantage in the market, and stability in the long run (within its industry) [17]. BItechnologies provide historical, current and predictive views of business operations. Common functions of businessintelligence technologies are reporting, online analytical processing, analytics, data mining, process mining, complex eventprocessing, business performance management, benchmarking, text mining, predictive analytics and prescriptive analytics.Business intelligence aims to support better business decision-making, thus a BI system can be called a decision supportsystem (DSS) [5].BI tools are a type of application software designed to retrieve, analyze and report data. The toolsgenerally read data that have been previously stored, often, though not necessarily, in a data warehouse. BI has experiencedhigh growth and BI technologies have gained lot of popularity. According to Hancock & Toren, 2006, BI is “a set ofconcepts, methods, and technologies for turning separated data in an organization into useful information in order toimprove business performance”[15]. In a Business Intelligence environment, data from various sources are extracted,transformed and loaded (ETL) into an Enterprise Data Warehouse (EDW) and, from EDW, they are used for generation ofreports across the organization.BI process and its various stages are shown in Fig 1. Data quality plays an important andcritical role in BI success, since poor data quality can affect business decisions at all levels of the organization and suchpoor quality data also hinders the growth of organization.
The fact is that the business runs on data; it acts as fuel for the corporate industry engine. A company cannot understand itscustomers, suppliers, competitors or its own people, processes, and company performance without good data. Therefore,business and IT should work together to ensure high-quality data [11

REVIEW OF LITERATURE
In a 1958 article, IBM researcher Hans Peter Luhn used the term business intelligence. He defined intelligence as: "theability to apprehend the interrelationships of presented facts in such a way as to guide action towards a desired goal. [16]"Business intelligence as it is understood today is said to have evolved from the decision support systems which began in the1960s and developed throughout the mid-1980s. DSS originated in the computer-aided models created to assist withdecision making and planning. From DSS, data warehouses, Executive Information Systems, OLAP and businessintelligence came into focus beginning in the late 80s.
Howard Dresner (1989) proposed "business intelligence" as an umbrella term to describe "concepts and methods toimprove business decision making by using fact-based computerized support systems.[9]" It was not until the late 1990s thatthis usage was widespread.
Stackowiak et al. (2007) defined Business intelligence as the process of taking large amounts of data, analyzing that data,and presenting a high-level set of reports that condense the essence of that data into the basis of business actions, enablingmanagement to make fundamental daily business decisions[20]. Cui et al. (2007) view BI as way and method of improvingbusiness performance by providing powerful assists for executive decision maker to enable them to have actionableinformation at hand[8]. BI tools are seen as technology that enables the efficiency of business operation by providing anincreased value to the enterprise information and hence the way this information is utilized.
Zeng et al. (2006) define BI as “The process of collection, treatment and diffusion of information that has an objective, thereduction of uncertainty in the making of all strategic decisions.” Experts describe Business intelligence as a “businessmanagement term used to describe applications and technologies which are used to gather, provide access to analyze dataand information about an enterprise, in order to help them make better informed business decisions.”[24]
Tvrdikova (2007) describes the basic characteristic for BI tool is that it is ability to collect data from heterogeneous source,to possess advance analytical methods, and the ability to support multi user‟s demands. Zeng et al. (2006)categorized BI technology based on the method of information delivery; reporting, statistical analysis, ad-hoc analysis andpredicative analysis.[21]
The concept of Business Intelligence (BI) is brought up by Gartner Group since 1996. It is defined as the application of aset of methodologies and technologies, such as J2EE, DOTNET, Web Services, XML, data warehouse, OLAP, DataMining, representation technologies, etc, to improve enterprise operation effectiveness, support management/decision toachieve competitive advantages.
Golfarelli et al. (2004) defined BI that includes effective data warehouse and also a reactive component capable ofmonitoring the time critical operational processes to allow tactical and operational decision-makers to tune their actionsaccording to the company strategy [14].
Gangadharan and Swamy (2004) define BI as the result of in-depth analysis of detailed business data, including databaseand application technologies, as well as analysis practices. They widen the definition of BI as technically much broadertools, that includes potentially encompassing knowledge management, enterprise resource planning, decision supportsystems and data mining.[13]
Berson et.al (2002); Curt Hall (1999) BI includes several software for Extraction, Transformation and Loading (ETL), datawarehousing, database query and reporting, OLAP, data analysis, data mining and visualization[2].

TYPES OF DATA PROBLEMS
There are many types of data problems that effect business and its processes; they can generally be divided into twocategories: defective data and inconsistent data [11]. Most of the companies don‟t take the necessary steps to figure out andaddress these issues until it has a bad impact on their business.
A. Defective Data
(i)Types of Data Defects
There are many types of data defects. Inaccurate, incomplete, unavailable or obsolete data is defined as defective data. Forexample, manually entered data is often full of spelling errors, typographic errors, and perhaps falsified. Sometimes peopleneglect to fill in all the fields or enter correct data in the wrong fields. Data defects can also appear when a system is movedfrom one platform to another, an old application is replaced with a new one, or data is moved from one application toanother on an automated and scheduled basis.
Fixing Data Defects
The problem with defective data is that it is difficult to find out once it‟s entered into the system. The best way to preventdefective data is to prevent it from being entered in the first place, so for this the businesses should invest in systems thatvalidate and fix data at the source when it‟s entered into the system or moved between systems through an applicationinterface. To fix defective data already in the system, the business must invest in data profiling and cleansing tools tocleanse and validate data sets before they are loaded into data warehouses for further use in BI.
B. Inconsistent Data
(i) External Data
Inconsistency in data occurs over time as the data is either duplicates or out dated. For example, customer data degradesover time as people marry, divorce, die, move, or change their names, so, this results in inconsistent data.
(ii) Internal Data
Internally, companies fragment data to suit their needs, largely because of the way they are organized and also because theylack centralized data management. Each analyst, department, or division sees the world through its own lens and thusdefines and models data differently.

CHARACTERIZING DATA QUALITY IN BI
Data quality can be measured and quantified according to various parameters. Previous work provides differentclassifications of the data quality dimensions. By analyzing these classifications, it is possible to define a basic set of dataquality dimensions including accuracy, completeness, consistency, timeliness, interpretability, and accessibility, whichrepresents the dimensions considered by the majority of the authors [19].
BI-specific data quality dimensions
Completeness: deals with to ensure “Is all the requisite information available? Are some data values missing, or in anunusable state?
Completeness measures to which extent data should have been recorded in a table is effectively present and also checkswhether it is according to the warehouse specifications. We refer to vertical incompleteness when we measure completenessof data in a column. Horizontal incompleteness refers instead to the quantity or percentage of entire tuples. Completenesscan be modeled as an extensional or an intensional metadata, and at different levels of granularity. Extensionally,completeness for cells is expressed as a binary value (i.e., true/false). Intensional measure of incompleteness refers to asituation where a rule is stored rather than a measure. We can define rules as functions over the dataset that identify a set oftuples, and for these tuples, the rules define a completeness measure in terms of its percentage.
Consistency: Consistency denotes the uniformity of the information in a given table. Syntactic consistency refers touniformity in the data format for a specific field. This is typically something that is detected and corrected at data cleaningtime (e.g. via normalization). Semantic consistency refers to the satisfaction of semantic rules defined over a set of dataitems.
Validity: Validity refers to the correctness of data.
Conformity: Conformity refers to expectations that data values conform to specified formats? If yes, do all the valuesconform to those formats?
Accuracy: Do the data objects accurately represent the “real world” values they are expected to model? Incorrect spellingsfields and outdated data can impact operational and analytical BI applications. Confidence describes the perceived accuracyand precision of the data, or the degree of trust that the data present in a table or set of tables is accurate.
Integrity: What data is missing important relationship linkages? The inability to link related records together may actuallyintroduce duplication across the systems.
Some other data quality characteristics are Assurance, Delivery, Openness/ Transparency, Responsibility

CATEGORIES OF BUSINESS INTELLIGENCE
A. Types of business intelligence tools
The key general categories of business intelligence tools are [4]:
• Spreadsheets
• Reporting and querying software: tools that extract, sort, summarize, and present selected data
• OLAP: Online analytical processing
• Digital dashboards
• Data mining
• Data warehousing
• Decision engineering
• Process mining
• Business performance management
• Local information systems
Incremental and Iterative process followed in BI architecture is shown in Fig 2.
B. BI Delivery Mechanisms
Scheduled Reports: These reports are generated at regular intervals and provide the “as – is” information to business.These typically answer the question – „What happened?‟
Ad hoc/User Query Tools: With the help of the available BI tools, business can query the underlying data and makedecisions. These typically answer the question – „Why did it happen?‟
Dashboards: These are BI solutions where analysis and reporting tools are used to provide feedback on the achievement ofKPIs. They are also used to do ad-hoc reporting to improve organization‟s ability to make correct decisions. Dashboardscan be prepared on claim reporting and settlement time, sales service support team‟s response times, etc. These typicallyanswer the question „As a company, how are we performing?‟
Trend Analysis Reports: These business analytics solutions involve data mining to determine the historic behavior of agroup, or the performance of a grouping of people, risks or transaction types. Trend analysis reports can be prepared onsegment of customers by risk, types of claims by occurrence, etc.
Forecasting: This BI solution allows an insurer to attempt to figure out what will happen in the future based on evaluationof current and historic data. This typically answers the question –„What is going to happen next?‟
C. Key Features of BI Architecture
Listed below are some of the key features of a flexible and highly adaptive BI architecture:
• High quality data: Data presented to the business should be of high quality and consistent. Users from variousgroups should not end up arguing regarding whose data is correct. Everyone should get the same numbers whenquerying a metric.
• Data which is presented to business should be correct, complete and current.
• Data sources should be accurately identified. Business and IT should work together to find most suitable datasources.
• Data storage and maintenance – DBA‟s should establish processes, checks and balances for data storage andmaintenance.
• Selection of suitable ETL (Extract – Transform – Load) Tools: Depending on the organization‟s requirementappropriate ETL tools should be chosen.

AGILE BUSINESS INTELLEGENCE
Agile business intelligence is an approach that combines processes, methodologies, organizational structure, tools, andtechnologies that enable strategic, tactical, and operational decision-makers to be more flexible and more responsive to thefast pace of changes to business and regulatory requirements[6].
The future of agile business intelligence:
In the past, BI vendors and BI application developers focused on business and operational functionality and architecturalrobustness. BI vendors and developers now need to concentrate on technologies that are categorized as "agile" and refer tofour major subcategories of agility: automated, pervasive, unified and limitless. Each of these new technologies stands onits own and is independent of the others. Firstly, firms need to automate BI processes and steps as fully as possible toeliminate manual work and free up the human resources for analysis and other value-added tasks. Next, as BI initiativesattempt to bridge data and information silos, BI technology itself is not unified. Today, different BI tools address various BIuse cases, such as historical and predictive or batch based and real time BI applications. Next-generation BI brings all ofthem together in a unified platform. After automation and unification, companies should address pervasiveness. How?Make enterprise BI applications available wherever and whenever strategic, tactical, and operational decision-makers needto analyze information, make decisions and act. This category includes technologies such as mobile, cloud and embeddedBI. Finally, the next-generation BI to be able to face the challenges of the modern business world - a world that does not fitinto nice, neat models - it must operate on information without any borders or restrictions using technologies line inmemoryand other BI specific databases.

CONCLUSION
In summary, comprehensive data quality management is a necessity for BI success in the organization as data qualityrepresents a massive risk for the BI effort. BI processes have to incorporate steps for dealing with data quality issuesoriginating from source systems and assure data quality within BI processes. But data quality management is not only aboutmonitoring and fixing data quality issues. It is also about educating all types of BI users about data quality, data collectionfrom legacy systems and the BI environment including data changes while ETL processes. In this paper we have discusseddata quality characteristics, types of data errors, various stages of business intelligence that act as sources for data qualityissues. We have also covered the categories and key features of Business intelligence. Further research can be done onvarious BI tools available and exploring and analyzing which domain they are used, their technical specifications andrequirements, analyzing the most widely used BI tool. Another area of further research is mobile, cloud and embedded BI.

ACKNOWLEDGMENT
would like to thank Dr. Sawtantar Singh Khurmi, Professor, Bhai Maha Singh College of Engineering, Muktsar for his support, encouragement andvaluable comments. He guided me and offered invaluable detailed advices on grammar, organization, and the theme of the paper.

Figures at a glance
Figure 1 Figure 2
Figure 1 Figure 2

References