Overview

Data modeling in DJ can be done in several ways:

• Using the UI
• Using the API
• Any supported clients
• Using YAML files

Data Modeling Stages #

The typical flow for onboarding data models looks like this:

1. Create appropriate namespaces for organization.
2. Register tables as source nodes in DJ.
3. Create transform nodes if any additional, light-weight SQL transformations are necessary.
4. Create dimension nodes, depending on what dimensions are needed. These may be simple references to existing source nodes, if the dimensional modeling has already been done outside of DJ.
5. Link dimensions into your source, transform, or dimension nodes to build out DJ’s dimensional graph.
6. Create metric nodes, which are aggregate expressions on one or more fields from a source, transform, or dimension node.
7. Create cube nodes for bundling commonly used metrics and dimensions together (often used to facilitate materialization for quick access, like when powering analytical dashboards).

We’ll walk through this in more detail in the Data Modeling Tutorial.