Data integrity is critical to the validity of any BI solution. The old adage “Garbage In, Garbage Out”, (GIGO), holds true in the world of data warehousing. A data warehouse with data that has no integrity renders any reporting and analysis effectively useless. The discipline of data governance puts in place a set of processes that ensures that data can be trusted and that there is accountability for data quality issues.
Data governance encompasses data quality and assurance, data security and metadata management and defines an organisation’s method of proper handling of data. One often overlooked and important part of data governance is metadata management. Being able to answer questions about different values appearing on two reports for the same measure, (for example a Year-on-Year Profit % comparison), can often take several days and cost the inefficient use of staff resources. These sort of questions can be answered by carefully managed metadata that can provide a piece of data’s lineage.
Data lineage allows the tracing back of a piece of data’s source through different levels, (eg. database views, datamart etc.). Providing a user with the capability to understand where the data comes from and what happens to it before it is presented in a report can provide trust in the reporting solution.