Do we really need a data model? As a data modeler, I hear this comment more than I would like. It has nothing to do with newer Agile and shorter time to delivery methodologies. I have heard it for the past 25 years I have lived in the relations DBMS world. I thought I would share my top ten reasons I talk about when faced with develop teams posing the above question.
5 Reasons to build a logical data model
- It’s a roadmap. –You would not set off on your vacation without a roadmap; be it a Google Maps map or GPS enabled navigation. The process of transforming business rules into a data model creates the starting point and path for the development of the database.
- Take a picture. It lasts longer. – It can be argued that the caveman was the first data modeler. He learned that he could use pictures to communicate. Fast forward to today’s high tech world. Database diagrams distill the complexities of a soon-to-be application into a graphic that is easily understood.
- We need some organization. – The logical data model works like a closet organization system. We take an unordered assortment of business facts and store them in an orderly manner for future search and retrieval.
- Let’s tell a story. –Spreadsheets and word processed tables can record business facts. However, telling the story of the business need is cumbersome when cobbled together in MS Word and MS Excel documents. The data model diagram eads like a book. It allows us to tell the story of the data on a single page.
- Bridging the business and IT gap. – The logical model speaks the business’ language in a format that IT understands. It is a tool that transforms business requirements into structures that the IT world uses to start the application database development process.
5 Reasons to build a physical data model
- Provide a blueprint for developers. – The logical model does not speak the language of the application development team. The physical data model transforms the business objects into database objects. This serves as the blueprint for all interactions with the database.
- Optimizing the data for the consumers. –The physical model converts business speak into technical structures and objects for the application development and deployment world. Beyond IT, consumers are anyone with a reporting or analytical tool connected to the database.
- Playing nice with the DBMS. – The logical model needs to be optimized to capitalize on the database’s efficiencies. It must also live with the DBMS deficiencies and limitations. Today’s data modeling tools provide a platform to perform this manipulation.
- Ease of generating the database. – Everyone likes a tool that makes life easier. Generating the database DDL is a common feature of data modeling tools. The DBA supplements the data modeler’s design with database characteristics to make this repeatable process smoother and easier to manage.
- Deploying the metadata to the masses. – Metadata has accumulated through the data design process. Physical model objects become the database where the metadata can be more easily deployed. Physical data model is a source for end user reporting and analytical tool metadata.
Summing it up
The above items are just a sampling of reasons to create a data model. Every data professional has their list of reasons to model. Some are certainly variations of these reasons. Others are unique to the individual or enterprise. Take a few minutes and write down your points. Learn to speak to them in your everyday work. It boils down to effectively communicating the role of the data professional to the non-data professional world.