Deliver Instance¶

Creating curated data marts, and adding endpoints to a Deliver instance.

In this module: creating a semantic model · role-playing dimensions · date hierarchy

What is a Deliver instance?¶

A Deliver instance is where prepared data becomes business-ready semantic models for analytics tools, so users always get consistent, governed, easy-to-understand outputs. It is the final layer in the architecture, and where you publish curated datasets.

In the Portal, go to Data Estate → Instances, choose Add new instance → Add deliver instance, and fill in the instance name, description and log retention. The critical step is configuring endpoints — selecting the types you need (Power BI, Tableau, Qlik, CSV, Tabular) directly in the setup form.

The process is deliberately low-code and form-guided, which centralises endpoint management and allows consistent deployment of semantic models to several BI environments at once. Once saved, the Deliver instance is immediately available in the Desktop application for modelling.

The step-by-step creation is covered in the Portal module. Reference: Add Deliver Instances.

The semantic model¶

A semantic model is a business-friendly abstraction over your technical data infrastructure. It translates complex structures into meaningful business concepts: instead of exposing raw tables, columns and joins, users see familiar names like "Monthly Sales" or "Customer Lifetime Value" — not FactSales.Amount or dimCustomer.ID.

Why it matters:

• Business-centric access. Removes technical barriers, letting business users work with data in their own language.
• One version of truth. Finance, sales and operations all use consistent definitions, calculations and dimensions, preventing "multiple versions of the truth".
• Data democratization. Reduces dependence on IT — teams explore, report and use trusted data products independently.
• Faster decisions. Consistent, findable data means faster reporting and quicker action.
• Governance built in. Rules and definitions are managed centrally, improving data quality, lineage and compliance.

Add the Order Date table to the sales model¶

1. From MDW, drag the Date table onto the Sales Deliver instance.
2. Name it Order Date.
3. Select DateValue and rename the field to Date.
4. Select YearMonthName and rename the field to Month.
5. Select YearQuarterName and rename the field to Quarter.

Add the Ship Date table — a role-playing dimension¶

A role-playing dimension is one where the same dimension table plays a different role by relating to a different foreign key on the fact table. Here, the date dimensions hold identical data, but relating to ShipDate instead of OrderDate filters and shows different results.

1. From MDW, drag the Date table onto the Sales Deliver instance again.
2. Name it Ship Date.
3. Rename DateValue to Date, YearMonthName to Month, YearQuarterName to Quarter.

Add the Customer table to the sales model¶

1. From MDW, drag the Customer table onto the Sales Deliver instance.
2. Name it Customer.
3. Select these fields: AccountNumber, CustomerName, CustomerType, PhoneNumber, EmailAddress.

Add the Product table to the sales model¶

1. From MDW, drag the Product table onto the Sales Deliver instance.
2. Name it Product.
3. Select these fields: ProductModelName, Color, ProductName, ProductCategoryName, ProductSubcategoryName.

Reference: Create a Role-Playing Dimension · Deliver instance relations · Measures · Custom measures

Next in the course¶

Semantic Model Endpoints · Data Cleansing