You are here:  Home  >  Data Blogs | Information From Enterprise Leaders  >  Current Article

Data Modeling: Why Not 3D?

By   /  October 19, 2016  /  2 Comments

Click here to learn more about author Thomas Frisendal.

Data Models are meant to communicate structure of information to recipients, who need to understand it. But who are these recipients? In fact, they are human beings. Given that, we should choose our means of communication to be the most effective that we can find. Other professionals have realized this. So what have they done?

They listened to the cognitive scientists and psychologists, and learned that vision is our primary sense (80 % of our cognition is driven from it). Here are a couple of inspiring examples of findings about human perception. First an example of visual reading:

“Rinadeg Oedrr

Aoccdrnig to a rscarhee at Cigdmabre Uinervtisy, it deosn t mtetar in waht oredr the ltteers in a wrod are, the olny iprmoatnt tihng is taht the fnst and lsat ltteer be at the rghit pclae. The rset can be a taotl mses and you can sit11 raed it wouthit porbelm. Tihs is bcuseae the huamn mnid deos not raed ervey lteter by istlef, but the wrod as a wlohe.”

(From: Interactive Data Visualization: Foundation, Techniques and Applications. Second Edition, by M.O. Ward, G. Grinstein, D. Keim, CRC Press 2015)

And then visual recognition:


(Source: Henri Rousseau [Public domain], via Wikimedia Commons)

Just a few visual clues make the difference: You’d better run, now!

Visual cognition is so powerful that we simply must exploit it.

I have an ultimate vision. Let me share it with you using the contemporary paradigm of an Agile user story:

As a Data Modeler I want to be able to use the same kind of 3D tools that other architects use so that I can easily explore and verify my design together with my stakeholders.

This means that all processes should be supported by 3D visualization tools, and that model prototypes should be available in 3D print. Architects (of buildings), designers (of things) and modelers of molecules, among others, do this everyday, now.

They know by experience that visualization of complex structures is what works for discussing, brainstorming, refining and even convincing and selling elaborate solutions. So that is what they do.

The power of visualization is backed up by solid cognitive psychology, and in our own data-centered neighborhood, advanced Analytics is more than saturated with expressive visuals.

Are data models visually expressive?

We are, at best, equipped with 2D computer aided modeling software, and we produce nice engineering diagrams such as this:


ERD diagrams are not truly business-friendly (to be diplomatic). Just like the building architects’ CAD drawn floor plans:


The molecular biologists went 3D a long time ago with their balls and sticks models:


(By Ben Mills – Own work, Public Domain, https://commons.wikimedia.org/w/index.php?curid=6875460)

They are printing them in 3D now: https://youtu.be/RvQ6R_op128

And the architects (and designers in general) managed to go 3D in a responsive manner:


(From augmented reality specialist Augment: http://www.augment.com/wp-content/uploads/2015/06/architecture-4.png)

3D printed building models, here is an example: https://youtu.be/NbXM8xcz8PA

Can we do a similar thing? And create the same kind of fascination from our “readers”?

I say: Yes, we can! Why not?

Structure can be represented as directed graphs; look at concept models, concept maps or even mind maps. Graphs are visual. There is a short way from the molecule model to a 3D Data Model represented as a graph.

And here is what you can do today for browsing a graph and speak directly to the human cognitive capabilities – remember to come back here after watching 🙂

Bloomberg Beta Quid for Investors and VCs: https://youtu.be/p2F5fiLP1O4

And why not go all the way? A 3D printed Data Model? Sure thing.

Take up the challenge Data Modeling tool providers! The graph visualization below is PhD work done by Peter J. Radics in a project called OSNAP – The Open Semantic Network Analysis Platform:


Well, Data Models are Semantic Networks too. Don’t miss this train, go 3D now, please.

There is no reason denying it. You, the reader, is also a visionary! We all are. Visualize to get mesmerizing Data Models and data. Graph representations is the Excel of information structures and are immediately fascinating to the human mind.

In up and coming blogs in this space I will inspect more mundane and classic data modeling challenges. But I will always have the vision of a 3D Data Model in my head. Until some software vendor or other makes my day and launches the solution.

If anybody has seen a visual 3D Data Modeling Tool, please let me know!

About the author

Thomas Frisendal, Owner and database consultant for TF Informatik Thomas Frisendal is an experienced database consultant with more than 30 years on the IT vendor side and as an independent consultant. He has worked with databases and data modeling since the late 70s; since 1995 primarily on data warehouse projects. He has a strong urge to visualize everything as graphs - even datamodels. His area of excellence lies within the art of turning data into information and knowledge. A fresh approach to information-driven analysis and design, which is "New Nordic" in the sense that it represents the Nordic values such as superior quality, functionality, reliability and innovation. In short: New ways of communicating the structure and meaning of the business context. Thomas is an active writer (see more about his two books on the website referenced below) and speaker. He works as a Data Architect and Modeler out of Copenhagen, Denmark. He owns the consultancy firm TF Informatik, registered in Denmark.

  • Joe Celko

    The ERD diagram you showed was really awful. If anybody submitted this in one of my classes, I would flunk them. Being someone’s supervisor is not an attribute; it is a relationship among employees. Identifiers are never numeric (what math do you do with them? What is the square root of your credit card number?) Why is department manager not an employee in this model?

    I know what you are trying to do by showing an ugly complicated looking diagram, but you could have made your point with a valid diagram instead.

    • Thomas Frisendal

      Thanks Joe,
      The flaws and faults that you highlight are indeed not best practise. But I have seen them in real data models out there. Doing data warehouse work for the last 20 years makes you a bit cynical. That is what I act on, trying to revive proper data modeling with new technologies.

You might also like...

Property Graphs: The Swiss Army Knife of Data Modeling

Read More →