Warning: Creating default object from empty value in /home/johnnz/public_html/wp-content/themes/simplicity/functions/admin-hooks.php on line 160

The Universal Data Compass

Fuzzy thinking is spreading like a fog across the landscapes of data management and data quality.  Instead of extracting themselves from this fog, many data analysts actually convince themselves that, far from being lost in the mists and haze, they are actually finding enlightenment in a new and unexplored world of complexity.  Few ever emerge from these mists; none emerge enlightened.

It is sad to see so many lost souls.  So, to help them escape their doomed wanderings I want to share with them a powerful navigation tool, the Universal Data Compass.

Like the magnetic compass that navigators have used to plot safe courses for thousands of years, the Universal Data Compass also has Four Cardinal Points.

Four Cardinal Points

Data in any enterprise, no matter what its size or commercial sector, falls into four categories, which are Master Data, Transactional Data, Domain Data and Metadata.The Data Compass helps enterprises plot a save passage to Data Quality.

Master Data: This represents  those entities which, when linked through commercial transactions, create value or generate income for the enterprise; these entities are Party, Product, Location and Asset.

Transactional Data: This represents all of the commercial transactions that the enterprise might enter into with another Party, and include such things as sales, purchases, contracts, etc.

Domain Data: This represents sets of data whose values are used to control and validate the values of the attributes of Master and Transactional Data Entities.

Metadata: This is data that describes the elements, formats and structures of all of the Master, Transactional and Domain data entities in the enterprise.

Metadata Muddle

The thickest mists currently swirling in the data landscape surround the term ‘Metadata’. For many years this term had a very specific meaning, which was ‘data that described data’.

However, in recent times, the term has started to be incorrectly used in place of the simple term ‘data’. It’s not quite clear where this misuse first occurred. Possibly in the area of digital photography where, for some strange reason, the data that described digital images came to be termed ‘metadata’ instead of simply ‘data’.

More recently, the misuse has spread into the media and, more worryingly still, it now seems to be spreading among data analysts and even into those bodies responsible for defining data standards.

Orientation Rules

There are two further orientation rules that will help any analyst to safely differentiate between ‘data’ and ‘metadata’. However, in order to clearly explain these we need first to agree on the meaning of an ‘instance of an entity’.

If an enterprise has a data entity called Person, then an instance of this entity might be ‘John Owens’. Another instance could be ‘Fred Smith’.

In an engineering enterprise, an instance of the data entity Product might be ’25mm stainless steel nut’ or ’25mm x 125mm stainless steel bolt’.

So now, with our definition of an ‘instance of an entity’ clear, we can give two orientation rules that will prevent all data travellers and explorers getting lost:

  1. If the data you are looking at represents instances of Master, Transactional or Domain data entities, then the correct term for this data is simply ‘data’
  2. If the data you are looking at represents or refers to the elements, structures and formats of Master, Transactional or Domain data entities then the correct term for this is ‘metadata’.

In all circumstances, other than 2 above, data should always be referred to simply as ‘data’.


For more examples of how to differentiate between ‘data’ and ‘metadata’ read my posts What’s the Matter with ‘Meta’? and More on the Meta Data Muddle.

Click here to pick up a comprehensive Glossary of 160 Pragmatic Modeling Terms for FREE.

If you liked this post, please share it with colleagues and friends by clicking a social media button below. To follow IMM on Twitter or Facebook, click on a floating icon at the bottom right corner of this window so that we can stay connected!


2 Responses to “The Universal Data Compass”

  1. Jacob August 28, 2013 11:37 am #

    Hi John
    In media companies the data IS the photo or AV asset and the data describing the data is called metadata. I think the definition hold true, even for photos and other media data types

    • John Owens August 29, 2013 2:37 am #

      Hi Jacob

      Thanks for this input.

      A particular photograph that a media company might want to use or sell is not ‘data’. It is an ‘item of significance’ about which they need to hold data, i.e. a Data Entity. The data that describes Data Entities is not metadata, it is simply data. This is true whether the photograph is held as a negative, as a print or in digital format.

      It was probably the move to digital format that caused the confusion, as many data analysts made the mistake of thinking that the photograph had now become ‘data’. Not so, it is merely an entity held in digital format.


Leave a Reply to Jacob