Mastering Metadata!

The Metadata Malaise!

Metadata may not be the sexiest topic, nor may it be at the forefront of everyone’s mind. But it needs to be. !

The operational culture of the enterprise has gradually shifted towards a more business- centric model, one where business users and front line information consumers do not just desire better access to enterprise data assets, but need to have a comprehensive view of the organization. This means that data, as well as the metadata that describes it, must be shared, both at the developer level and the BI level. This is easier said than done.!

The importance of metadata within the data warehouse context cannot be understated. In many ways it is the backbone of enterprise data warehousing and information management. Metadata serves as a directory for everything that is in the data warehouse environment and most other data stores, providing detailed information about data and objects. In large organizations, metadata is typically dispersed and fragmented across departments and business units, and some of it resides in spreadsheets, text files and multimedia files. These files are rarely shared and accessible throughout the enterprise. !

You cannot share data effectively without managing the metadata resource. Here’s why. !

Metadata defines the structure of data in files and databases. Often this definition data is buried in program code, and only the program knows the precise data structure. That is not in and of itself the problem, as the program is quite able to use the data appropriately. The problem is that the metadata is meaningful to that program, but not to the human user or possibly even to any other program. If only one suite of business applications were in use throughout the enterprise, this might not be so constraining. However, almost no organizations have such a homogeneous software environment. !

Consider the following situation. Within a particular database the “Person” table is defined to consist of: Person-code, Title, First Name, Last Name, Job-Title, along with some data type information. Even if you are told that the data is from a database used by an HR application, there is not enough context to determine exactly what the data refers to. A specific row in the table might refer to a staff member, or a contractor, or a former employee. If this is the case, sharing this record would require an additional explanation as to what the reader is looking at. !

In practice this imprecision will be present for many tables defined in many databases across the business. As a consequence, when there are no implemented standards or methods for sharing data, it is practically impossible to ensure accuracy and consistency of reporting, analytics or decision making.!

The main reason for sharing data is to populate the diverse array of business intelligence and analytics applications within an organization. And yet application data is frequently

...