Why Model?

It is tempting today to model the data as you go. Schema-less or self-describing data helps a piece of the way. But you are still at risk of using bad assumptions and create flaws. For example not realizing that you are looking at a many-to-many relationship can be a difficult structural flaw to change two months down the line.
You can quote me for:
Data Modeling on the fly moulds like putty, but it sets like concrete
Thomas Frisendal
Stacks Image 122

In doubt about which Kind of Model is best for You? For most Use Cases, just Graph Your Data Model!

Since a data model is meant to communicate to people, you must be as visually mesmerizing as you can do it.

Using graphs to represent data models really changes the way we look at data models.
Stacks Image 67
Visualizing structure and meaning is the goal; because that focus greatly improves the communication to business stakeholders as well as to developers.

Furthermore this representation (the property graph model) is agnostic of the physical data store and its capabilities.

Data Modeling in the complete Picture

The visual approach to data modeling is described in detail in this book:
Stacks Image 105

Knowledge Graphs - Enabling Navigation in Complex Contexts

Enterprise Knowledge Graphs get a lot of attention these days. They are certainly very attractive applications of graph technology. Understand why by reading my blogpost over here at Dataversity: How Not to Get Lost in 2018 with Knowledge Graphs: Map, Graph, Go!
And don't forget to visit our Data Model History page for an overview of the important events and products on the database scene.
You may follow the sequence or explore the site as you wish: