Data models are not 'just documentation'

[GeorgeMcGeachie]

My friend Dirk Lerner is speaking at the IRM UK data management conference this week (it has a longer title, but I'll avoid using it today) on topics that are close to my heart, so I thought I'd share some related connections. Dirk has three sessions this week -

• Data, Data, Metadata – DNA of Data Architecture?


• From Conceptual to Physical Data Vault Data Model


• Send Bi-Temporal Data from Ground to Vault to the Stars

Why am I so interested in these sessions (especially the first two)? The answer is simple - he's discussing the combination of metadata and automation to improve the efficiency and efficacy of data modelling. This is a message I've been banging on about for years to anyone who will listen - data models are not 'just documentation', they're an essential part of application development.

There are tools out there to help you with automating the development of Data Warehouses, including Data Vaults. Some are commercial products which you can buy with services, others are part of a consultancy package (so you buy the service, and they use the tools to provide the service).

All of the consultancy examples I know of have extended a data modelling tool to provide the automation (they all use SAP PowerDesigner).

If you want to build your own automation, there is an initiative (led by Roelant Vos) to build a generic metadata interface for you to use when integrating tools - find out more at http://roelantvos.com/blog/a-collaboration-for-a-common-metadata-model/, and you can contribute to this effort at https://github.com/RoelantVos/Data_Warehouse_Automation_Metadata_Interface.