Metadata are data about data. This concerns data that aid in understanding and interpreting the meaning of recorded data in a specific context.

Metadata include not only information on the data themselves, what they mean, the format they are in, where they came from and the value that they can take, but also information on their mutual relationships and the means by which they can influence each other.

What does Metadata management means?

Metadata management or administration of metadata means targeted building and maintenance of a uniform, standardised repository (metadata warehouse) that contains the required information on data and relationships between them in a well-arranged format. This repository also makes it possible to input or update existing data. Information in the metadata warehouse is accessible in the same manner as “ordinary” data in a database. In addition, there is a range of tools that facilitate administration of the metadata warehouse and offer a clear visualization of the mutual relationships.

Uses of metadata

Typically both the IT department and business units of a company work with metadata. Business metadata contain contextual information on business concepts that are worked with in a given area (customer, employee, profit, costs, working period, contract, hierarchy of responsibilities). These metadata serve primarily to support qualified decision-making and the business user’s understanding of the data’s content. Technical metadata comprise contextual information of data sources, applications and tools (databases, ETL, OLAP, reporting, metamodels), and help IT workers to effectively maintain data sources and developers to develop ETL processes and reports. 

Metadata management ensures that data have a uniform meaning across an organisation and have a demonstrable origin. Without metadata management, there are always a certain circumspection and mistrust of outputs from the data warehouse. For example, typical uncertainties can appear thus:

• Is the “customer” actually what I understand by the term “customer”, or does this mean “payer”?
• Was “profitability” in this report calculated in the same way as in the report used by my subordinates?
• Are data transformed correctly in the ETL processes? Has this led to a loss of information somewhere or to imprecise calculations?

Properly processed metadata are also decisive for the quality design of ETL process and necessary for analyses related to data quality.

How to get started with Profinit?

Profinit offers analysis of the operation of the system for metadata management in your data warehouse, identification of potential problems and proposals for solving them. We can also design the introduction of a system for metadata management if there is not any solution for this area in your data warehouse. In the past we have assisted customers in these areas:

• introduction of a professional tool for metadata management
• assurance of metadata extraction from SQL scripts and stored procedures (this is not supported by most professional tools)
• preservation of metadata during migration or a change of technology in the data warehouse
• custom development of a complete tool for metadata management (metadata management, visualisation of data flows)