Presenting Data Models

picture by Virket Holding via Flickr

So you have prepared a data model. You’re proud of your work. Now how do you explain the model to others? You want them to understand your work and you want their feedback. We can think of three techniques for presenting data models.

Use a Data Modeling Tool

An obvious approach is to present a data model with a data modeling tool. For a large model you would jump from subject area to subject area during discussion. You may also use popup screens to show detail such as definitions.

  • Advantages. This technique is easy for the presenter. There is minimal preparation. The audience sees the current model which can evolve and change up to the meeting time. The model can be live and updated during a presentation.
  • Disadvantages. This technique is not viable for a large audience. It’s also not suitable for a business audience. Participants find it difficult to grasp a model of any complexity when it is presented via a tool. They are overwhelmed by the sheer quantity. They find it hard to track all the pieces of a model, what they mean, and how they fit together. Many participants will just tune out. It will appear that you didn’t prepare.

The bottom line is that you should not use a data modeling tool for a general audience. Only use this approach for a small group of persons who are knowledgeable about a model. Such a group will have a context to aid understanding.

Use Slides

Another obvious approach is to present a data model with PowerPoint slides. You use a tool to construct the data model. Then you copy and paste diagrams into the slides and add explanation.

  • Advantages. This is a good technique for an expert audience. They already understand data modeling and want you to quickly get to the important points. With slides you can abridge a model, suppress detail, and focus on the core. You can display diagrams, definitions, and explanation together. The slides become useful documentation.
  • Disadvantages. Slides can be overwhelming. They provide a peephole into a model and the audience can lose perspective. Slides are a static snapshot of a model that can be changing. It’s a hassle and work to prepare slides.

Slides are helpful for explaining a data model to an expert audience. You can quickly communicate important ideas. You can also use slides for a business audience but it’s more difficult. With a business audience, it’s best to strip down a model to its core and add lots of explanation.

Nominally Recreate a Model

A more subtle approach is to recreate the model live, in front of an audience. We came upon this technique some years ago. You prepare a model in advance and use it as a guide. You tell the audience that you already have a model and are creating it as a presentation technique. You then create a model in front of the audience.

  • Advantages. Audiences like this approach. They can ask questions about the evolving model and understand why it is structured a particular way. It’s less work for the presenter than preparing slides.
  • Disadvantages. You have to run a meeting and be proficient with modeling. That’s a high skill barrier. You must be sure to capture comments from the audience. Diagrams are only part of a proper data model – textual definitions, rationale, and explanation are also important.

This is a valuable approach, but we seldom see it being used. Many developers just don’t think about it. This is a great approach for explaining a model to a business audience. You pull them along as you construct the model. It’s also good for explaining a complex model to an expert audience.

In Conclusion

We use all three techniques in practice. Our choice depends on the audience and nature of the data model. This is consistent with our general approach to software development. We believe that developers should be fluent with multiple software techniques and adapt to meet business needs.

Leave a Reply

Your email address will not be published. Required fields are marked *