AbstractThe advancement oforganizations and companies results to the production of massive data fromdiverse ranges of the operational environments.
Most of the organizationsencounter difficulties on how to manage and make use of real-time data fordecision making. Owing to this, the organization was required to use inadequate information to generatereports and make decisions. In themidst of this upheaval, data warehouse was implemented to manage enormousvolume of data from multidimensional operational sources and integrate into asingle repository for easy accessibility. Data warehouse supports organizationsin decision making and enhance the competitive advantage. Despite of thevaluable benefit of data warehouse, organizations have to give emphasis to the featuresfor effective implementation. Based on the evidence from the literaturereview, both organizationaland technical factors need to beconsidered prior in the development process to attain its success.However, more research isrecommended to ensure the sustainable data warehouse in the organization.
Hence, this paper presents a detailed analysis on the significant elements for the effective implementation of datawarehouse in the organization. Keywords: Data warehouse, ETL, Data warehouseimplementation, Data mart Contents 1 INTRODUCTION.. 3 1.1 Data Warehouse Development. 3 1.2 Data Warehouse Development Approaches. 4 1.
3 Data Warehouse Implementation. 5 2 LITERATURE REVIEW… 5 3 RESULTS. 6 3.1 Organizational factors.
6 3.2 Technical factors. 7 4 DISCUSSION..
8 5 CONCLUSION.. 9 1 INTRODUCTIONA data warehouse (DW) is a pool of data produced tosupport decision making; it is also a repository of current and historical dataof potential interest to managers throughout the organization (Sharda et al,2014).
It is characterized as subject oriented, integrated, time varying andnon-volatile. Data warehouse accumulates and integrates data from varioussources into a single repository. Such repository enables businesses to gather,organize, interpret and use data to enhance transactional practices (Gupta andMumick, 2005). Moreover, data warehouse facilitates analytical processingactivities such as querying, reporting, data mining and Online AnalyticalProcessing (OLAP).
During the early 20th Century, most of thebusiness organizations were technologically advanced and generated large amountof data from their transactional processes. Thus, they faced complications on managinggrowing and fragmented data to producereal-time reports for decision making (Sharda et al, 2014). In this sense, mostof the organizations emphasized on the effective use of data for real-business decision support (Carr, 2004). By then, the concept of data warehouse was proposed as the technology which can ensuregathering of data from different sources of operations.
Data warehouseenables the organization to reduce the transaction cost, increaseseffectiveness and retain competitive advantage (Zeng, Chiang, and Yen, 2003).Moreover, it supports the organization to adapt to the current environment, learn from its pastexperiences and position itself for the future (Ganczarski, 2006). Thus,data warehouse helps organization to manage strategic and real time informationfor decision making (Cooper et al., 2000). 1.1Data Warehouse Development Data warehouse development is theprocess of defining, designing, testing, and implementing a data warehouse inan organization. Manning (1999) proposed a data warehouse as anarchitectural model which ensures the flow of data from the operational systemsand terminating in the decision support environments. Based on the architecture,Sharda et al, (2014) divided data ware into two types of architectures;three-tier architecture which contains operational systems (server), the datawarehouse, and the DSS/BI/BA engine (the application server) and the clients.
Also, two-tier architecture consists of data warehouse and DSS engine fromwhich both components run on the same hardware platform. Thus, it encountersdifficulties for data access in large data warehouses while; three-tier has acapability of reducing the resource constraints owing to the separation offunctions. Moreover, the development process needs support from the management such as settingreasonable time frames and budgets, and user’s involvement to enhance its success(Sharda et al, 2014). In this conception, data warehousing requires theintegration of various tasks, components and coordinated effort of severalindividuals (Kimball, 2006).Extraction Transformation and Load (ETL) is anintegral component of the data warehousing process which supports reading datafrom one or more databases, converting the extracted data from its previousform into the required form and put the data into the data warehouse. Additionally,a data warehouse requires online analytical processing (OLAP), data miningcapabilities, client side analysis tools which can handle gathering of datafrom different sources to users. Theoverall architecture of data warehouse is illustrated in Figure 1 which identifiesthe main architecture components and the data flow throughout the system. Figure1.
Architecture of a Data Warehouse (Humphries, Hawkins, & Dy, 1999) 1.2 Data Warehouse Development ApproachesDecisionsupport is the main goal for the advancement of data warehouse to business organizations.For the effective decision making, each organization must employ one of thedevelopment approaches according to the organization need. Sharda et al (2014) identified two approaches;first, top-down approach initiated by Bill Inmon.
It is also known as theenterprise-wide data warehouse (EDW) approach which uses traditional relationaldatabase tools such as entity-relationship diagrams (ERD) for the developmentneeds of an enterprise-wide data warehouse. Second, bottom-up (data mart) approachproposed by Ralph Kimball. It employs the dimensional modeling which startswith tables. A data mart is asubject-oriented or department-oriented data warehouse and is built one at atime. However,there is no approach which is supreme than the other since organizations aredifferent. Thus, an organization can decide on the approach according to userdemands, the enterprise’s business requirements, and the enterprise’s maturityin managing its data resources (Sharda et al, 2014).1.
3 Data Warehouse ImplementationImplementing a data warehouse is a challenging process;it requires much efforts, various resources and time. Also, the complexity isdue to the ability of enhancing standardized, enriched and integrated data forreal time information from several sources. Despite of competitive advantagesoffered by data warehouse to organizations, data warehouse projectsneed to be considered carefully to avoid risks which are more serious (Sharda et al,2014). In this sense, datawarehouse implementation also needs a planned effort for success (Goldstein, 2005).Accordingly, for an organization to maintain the data warehouse usefulness, effectiveimplementation has to be considered. Hence, this paper aims to identify the elements forsuccessful implementation of a data warehouse in the organization by conductinga systemic literature review.2 LITERATURE REVIEW The literature review anticipated to study datawarehouse implementation state in the organization, with an emphasis on thecorresponding factors for the effective implementation. Searching for materialswas enhanced by databases from Uppsala University as well as Google Scholar.
Thesetwo databases always ensure the availability of reliable numerous resources. Inorder to engage only the relevant literature, the author narrow searching from thetwo sources by identifying only the peer reviewed literature within the subject.Additionally, the newspaper articles and blog posts were excluded.The author uses the term “Data warehouseimplementation” to search for literature in all databases.
As a result, UppsalaUniversity database produces 219 results while GoogleScholar yields 215 results. Thus, the author investigates only literatures withrelevant titles in the first phase. Afterward,100 papers from both sources were identified for further review. In the second phase, the author repeated the processwith a more analyzing approach and browses the whole literature and finally concentratingon the abstract and conclusion to identify the relevance. By then, 50 paperswere selected to guide the last phase of the review.
Owing to time, in the final review phase; the authordecided to use only 22 literatures provided that all results were found by usingthe same search term from the databases. Next, after a critical review of theliterature, the following findings in the next section were identified.