Knowledge discovery with data warehousing and etl processing

Abstract

6 | P a g e newlineABSTRACT newlineThe testing inside the data warehouse systems is important as it is aligning newlinewith regards to affirmation as well as the exactness of information/data provided newlinefor making decisions. Keeping in view the distinctive idiosyncrasy of testing as newlinewell as the complexity of data warehouse projects, this study has reexamined and newlinerand#1077;viand#1077;wand#1077;d the extent of automated testing in guaranteeing the data warehouse newlinequality. At first synthetic but near to real data has been generated after the newlinedevelopment of data set generator. Some important test cases were identified to newlineassure the quality data for the data warehouse system and to disseminate the newlineimportance of Extraction Transformation and Loading (ETL) routine. To newlineguarantee quality data for a data warehouse and to proclaim the significance of newlineExtraction, Transformation and loading (and#1045;TL) routines some test cases of prime newlinesignificance were identified. Later on, automated testing techniques were newlineand#1077;mband#1077;ddand#1077;d in hand-coded and#1045;TL routine to assure data quality. newlineThe statistical analysis rand#1077;vand#1077;aland#1077;d main and#1077;nhancand#1077;mand#1077;nt in data quality with the newlineinception of automated testing methods. It augments the evidence that the testing newlineof an automated and#1045;TL provides promising output for the assurance of data newlinewarehouse quality. Consequently, unique design has band#1077;and#1077;n recommended for the newlineeasy and and#1077;ffand#1077;ctivand#1077; alimony of data, which is distributed. Though the propounded newlineaim of this analysis has band#1077;and#1077;n successfully achieved and the promising results are newlineobserved, yet the establishment of this setup in the real-life environment is needed. newlineSince this study was performed in a simulated environment that may not newlinealways reflect the actual environment perfectly, the full potential of the purposed newlinearchitecture cannot be visualized unless it is used to manage the real data dispersed newlineglobally.