Cognite Training on Flexible Data Modelling (Hands-on)

I’m also keen on learning how to best use the UI within CDF to do data modelling. Much of what is in the documentation is all based on YAML files and programmatically configuring everything. I’m much more of a visual person and I suspect many find it simpler to model their data in a more WYSIWYG fashion similar to many graphdb products like Neo4J. Even the basic lift station and pump example (which barely scratches the surface) is pretty intense to keep track of in your head, especially when you have to intentionally think about spaces, containers, instances, etc. If you had a way to visually convey the idea of spaces and containers as schemas and maybe even Right Click->New Node and then walk through a simple wizard that with a few choices would create all of the backend magic, that would be a much better user experience. In fact, what would be fantastic is to take the lift station/pump example and show how that is done within the Data Modelling section in CDF in a YouTube video. That would be a great starting point

Hi @mheinze57 and thanks for these comments.

As of right now, due to significant amounts of feedback and observations, we are in the aftermath of shifting our perspectives for how we recommend approaching the construction of objects and their relationships in data modeling. 

First of all, there are some changes already in the UI to align what you see in the “Model your data” user interface with how the documentation describes building a graph in DM (meaning, you can review container and view designs in the UI, and we’re using that language). There are more coming... We’re iterating on the data modeling UI throughout the first half of the year to re-simplify the creation and maintenance of/consumption of data from interconnected data models, and the resulting graph(s).

As you have observed, the short-term of our shift has been to take a bit of a step backwards, and fully exposing the components of the data models in the DM UI. We wanted to encourage “getting to know the innards” of data models since how you build those innards help determine the resulting performance and behavior when your users start consuming data from the graph you’ve built.

That said, your insights are in line with what others have commented over the past few months, and we _are_ listening. We’re also working hard towards an opinionated workflow when building models and connecting them into a graph that will perform reasonably, and will take your insights into account as we work with designs and features over the next several releases of CDF.