Understanding Data Modelling

Data modelling is one of those areas I haven’t really tried to learn to much about. Or should I say: I’ve tried not to learn too much about. It just sounds complicated. Here’s what I knew before today:

1. You have some data. It can be customer contact details, purchasing transactions, patient treatment information, whatever. It’s stored in big databases — very unfriendly.
2. If you want to take a look at this data to get some answers, you have to pull it out. Make it easier to work with, summarize things.
3. In order to do that, you have to first understand what the data means, how it relates to other pieces of data. So write it down. Create a model:
   1. First there’s a logical model. You could write this on paper, or a whiteboard. You can get more detailed later using software.
   2. Then there’s a physical model. This is where you use the concepts you’ve drawn out to create a data store and relationships.
4. The fourth thing I knew is that I really don’t need to know the details!

But it’s important for business people, like me, to be involved in modeling the data we work with. The more we understand, the better. Initially I thought there could only be one type of model. There’s data, there’s a model. 1 to 1. No. Wrong. You can model things in lots of different ways, and it depends on:

• How much stuff you want to include in your model (scope)
• How generic you want your model to be (abstraction)
• If you want to model the current state or a proposed state (time)
• How you’re modelling — whiteboard, flowchart, software, physical model

Steve Hoberman wrote a great article that compares data models to camera settings. It’s an illuminating way to explore the subject.

The one point I don’t understand is the Filter/Function aspect … not sure why the Application point of view should be different from the Business point of view, or for that matter how the camera lens filter analogy carries over. Anyone want to explain that to me?

-IainR
Sr. Marketing Manager – The BI Builders