Meta data is often talked about in organisations and consequently often written about, but there is a common belief that it is hardly ever implemented - but is this true? Let's explore.
Meta data is certainly a popular tick-list item on tender responses for the simple reason that most, if not all, software tools used in the business intelligence (BI) arena cannot work without meta data.
Unfortunately, the belief that it is seldom implemented into organisations stems from the fact that the term is often loosely defined. Consequently, many organisations do not have an understanding of meta data and the value it can add to the business.
Put simply, meta data is in fact data about data! It describes the how, the when and the who of a particular set of data that was collected and formatted, and is essential for understanding the information that has been captured and stored.
Within the BI industry, meta data refers to all the information that is left over when the actual data itself has been dealt with. The actual data would refer to something such as the customer name, but the meta data would refer to the 'connection information' that leads to the customer's data.
As such, an organisation that has a BI system driving its information strategy has some sort of meta data somewhere, perhaps without even realising it.
Making sense of data
In fact, without even lifting a finger, most organisations probably have the following information stored somewhere:
* Technical table and column definitions - these are in the DBMS catalogues, and may even include descriptive information in the form of comments.
* Connection details - where all source and target data resides. This includes what type of DBMS the data is in and layouts of flat files.
* Transformations - if the use of an ETL tool is undertaken. It is keeping track in its internal meta data of how data is transformed.
* Report business definitions - the reporting tool's definition of the report probably works off meta data. What's more is that report elements often are described in business terms quite well within the reporting tool's meta data.
Within the BI industry, meta data refers to all the information that is left over when the actual data itself has been dealt with.
Cor Winckler is a technical director for the PBT Group.
The organisation may also have some unstructured data that it is unaware of. This unstructured data can include things like documentation, Web pages and training material. Be it structured or unstructured, what is important for organisations to note is that they need to make sense of and control this meta data for the data to work for the organisation. Why? Because of the following:
* Meta data is key to ensuring resources survive and continue to be accessible for the future.
* It aids in the identification and retrieval of essential information.
* It provides the vital link between the information creator and the information user.
Knowing where all this data resides and having the ability to access it, is essential to a comprehensive data and information strategy, contributing to efficient decision-making and overall business processes.
Making sense of this data can mean investing in a specific tool that helps to exchange the meta data and make it centrally available to the organisation or, it could simply mean making a list or register of where all the data resides - for easy reference. Whichever way businesses go about it, it is important to get their finger on the 'pulse' of the organisation.
Contrary to common belief, it is evident that meta data resides in and is implemented into organisations. As such, the question becomes not whether an organisation has meta data, but rather, where they have meta data. And this is referred to as the process of finding the elusive meta data within the organisation's meta data, and it's a tongue twister, believe me. It's called Meta Meta Data Data!
* Cor Winckler is a technical director for the PBT Group.
Share