Model any domain entity as a mesh node type — then master the analytical cube. Two parts: the generic foundations of modelling, and the data cube built end to end.
This is a hands-on course. You don't just read about data modelling — you run it. Every concept has a live cell you execute in the page, and every lesson ends with a small piece of code or data you write, with its checks already there and red until you drive them green.
Who this is for
Anyone who has to turn a real-world thing — a claim, an order, a task, a balance sheet, a portfolio — into a model a computer can hold, validate, total, pivot and chart without lying to you. No prior MeshWeaver experience required. You should be comfortable reading a little C#.
What you'll be able to do at the end
Part 1 — Foundations (model any entity):
- Model a business entity as a node type: atomic properties, and a content record that never duplicates node metadata
- Annotate a model with the framework's real attributes —
[Dimension],[IdentityProperty],[NotVisible],[MeshNode] - Format numbers correctly — money with
[DisplayFormat], a rate as a percent with[Percentage] - Validate model data —
[Required],[Range],[RegularExpression]and[Percentage], run by one engine - Wire foreign keys — references that pick, link and render
Part 2 — Data Cubes (the Firnwald model):
- Decompose a real statement into atomic positions and dimensions
- Choose a sign convention so that summing always gives the right answer
- Model computed values as data, evaluate with business-rules scopes, and slice & dice into pivots and charts
- Recognize a cube-shaped domain from its description and propose the model
How each lesson works
Every lesson has the same three bands:
| Band | What it is |
|---|---|
| 📖 Theory | The idea, kept short — and grounded in real framework source. |
| ⚡ Experience it | A live code cell. It runs when the page loads — read the code, read the result, then change it in your head and predict what happens. |
| ✍️ Your turn | A small exercise: fill in a record, an attribute, or a method body. A worked solution is always one click away. |
Set up your scratchpad (2 minutes)
The lesson pages are read-only, so to run your own code you need a place to write. Two options:
- Ask the Coach. The Data Modelling Coach is an AI agent that lives in this course. Paste your attempt and say "run this and tell me if it's right." This is the fastest path.
- Your own page. Create a
Markdownnode in your own space, paste an⚡ Experience itcell into it, and edit freely — it runs the same way these do.
The worked examples
Part 1 models an insurance claim — a plain entity, so the modelling ideas stand on their own. Part 2 is built around Firnwald Vorsorge, a fictional Swiss pension fund; the complete, narrated build lives in Data Cubes, and the full source is on GitHub.
Ready? Start with Foundations 1 — Node types & properties.
Contents
Part 1 — Foundations — models a plain business entity, an insurance claim; Part 2 — Data Cubes — builds the Firnwald Vorsorge balance sheet one node at a time.