Data Governance describes the practices and processes organizations use to manage the access, use, quality and security of an organizations data assets.
The data-driven business era has seen a rapid rise in the value of organization’s data resources. This has led to greater interest in leveraging Data Governance to ensure that best practices are followed when handling data assets.
Inconsistencies in data definitions or terms can cause data-driven organizations a number of problems, including:
— Loss of productivity
— Bottlenecks and slower time-to-market
Such problems make capitalizing on or reacting to market disruption difficult. Data Governance can help eliminate these problems and help make organizations more agile and responsive to change.
Data Modeling Tips for Data Governance
It’s important to use high-quality data for a data governance initiative to produce viable results. Data quality is directly impacted by the quality of the underlying data models.
Simple data models that just provide high-level database diagrams are not very useful for improving data governance.
Collaboration is an integral component of a data governance program. Using data consistently across multiple internal departments requires a collaborative effort when creating the data definitions used throughout the organization.
Proper collaboration around the data models used for the governance initiative will save time and aid productivity when engaged in data governance tasks.
Add Metadata to Data Models
The first modeling tip we will look at is to use metadata to extend and enhance data model entities. Adding business-related metadata to basic data models makes them more useful and understandable to the diverse audiences that will use them.
Taking the time upfront to include metadata pays long-term dividends by creating better models. Following are some items to include in data models.
- Extended definitions and notes that appear in data models help everyone understand the entities described in the models. Consider using a specific pattern when creating definitions to enhance data models and don’t rely on naming standards to convey all meaningful information.
- Data steward information should be part of data models to provide relevant information on who should be contacted if concerns arise. Identifying the source of a data definition enables issues to be resolved quickly and directly.
- Data privacy and sensitivity classifications are necessary to allow data models to be reviewed and discussed by various entities in the organization. Everyone involved with the data model needs to maintain an emphasis on data security and correctly handling sensitive information. Securing sensitive data is one of the goals of a data governance program and will be improved with informative metadata.
Include Data Security Requirements in Data Models
Security requirements are vitally important when creating viable data models. The need to secure sensitive data and comply with regulatory guidelines should be incorporated into data models to ensure everything is in place before implementation.
- Logical data models should include business security requirements such as those needed for encrypting, masking, and accessing data objects. This includes defining which business roles can view unencrypted or masked data.
- Physical data models are used by DBAs when implementing the models in databases. Technical security requirements need to be defined in physical models so data analysts can determine how data can be used, how it is masked, and how it needs to be protected.
More descriptive and informative data models contribute to data governance efforts and foster better communication throughout an enterprise. They increase the value of data resources by making them easier to use when addressing business requirements.