When you do your data profiling, this is the time to make sure all the metadata matches the data it is connected to. Data is useless unless its metadata describes it, so that it can be found and accessed when needed. It’s crucially important, for example, in data warehousing, where a business requires specific analysis and reports.
Types of Metadata
Every type of business needs different kinds of information. For a shoe manufacturer for example, it’s important to know quantities of particular shoe sizes that might be required. A photo library may have to know the date on which each of its pictures was taken, where, with what camera, and by whom.
So as well as providing a description, our metadata needs to be:
- Structural, to follow rules about how things are stored and how they can be grouped together
- Administrative, to follow the rules that help us manage the data, such as who created it and when.
Metadata allows us to retain and access the data we need under any criteria, search for what we want at any time and distribute it to where it is going to be of value.
Some businesses aim for ISO 23081, the international standard that defines the metadata needed to manage records. It helps data managers achieve quality metadata, and maintain it.
Don’t Neglect Metadata!
Metadata helps us to recognise what is relevant, put data into categories and groups and produce lists and reports. No IT department can afford to neglect metadata as part of its data quality initiatives.