Posted in Information Technology & Systems, Total Reads: 397
Definition: Data Quality Audit
Data quality issues are frequently widespread and start in the source frameworks, their applications, and operational procedures. Most of these issues are the immediate result of inadequate warehouse administration.
Keeping in mind the end goal to combat these issues, data warehouse architects must comprehend their source information. This understanding can originate from data profiling. In any case, despite the fact that profiling procedures are important, they're still exploratory. They leave it over to the analyst to see how the data profile fits the business prerequisites. That is the place a business principles based review can be valuable, if not critical.
The Data Quality Audit comprises of two sections. With the Data Analyzer, the profiling instrument, a precise outline of the state of the data is created, and an assessment is made based upon already chosen data quality basis. The procedures are also examined, in which ideally, the DQ model is settled by business rules. Together with the pioneers of the specialist departments involved, solid requests on information are resolved in an initial workshop. This data gives an idea regarding the quality standard needed. The results of the examination towards the end of the operations as received as a comprehensive documentation. Specific attention is paid to possible inconsistencies between the quality of the information itself, and the quality needed by the procedures. This individual analysis is used as a premise for development suggestions.
The data quality audit is a business principles based methodology that uses standard deviation to recognize variability in test outcomes. We look at the probability (confidence levels of 95 to 99 percent) of invalid information qualities occurring in sections or fields after business principles are connected against test information. A business principle may be a basic check, for example, "Account_Aggregate_Balance must be more than zero," yet it might also incorporate guidelines that guarantee its accuracy. For instance:
In the event that "Account_Aggregate_Balance" is > 0
Also, "Account_Type" = "Credit"
Also, "Account_Open" is > 30 days
Then, at that point "Account_Interest_Fee" MUST BE > 0