Power BI

AnalyticsCreator generates and maintains semantic-model definitions for Microsoft Power BI from the metadata and business logic defined in the AnalyticsCreator project.

For Power BI Project workflows, AnalyticsCreator can generate the semantic model and measures as native TMDL. These generated definitions can be versioned in Git, reviewed as text-based changes, and used in a configured Power BI or Microsoft Fabric CI/CD workflow.

Overview

AnalyticsCreator separates analytical model design from its technical implementation. Dimensions, facts, relationships, measures, hierarchies and other model definitions are maintained as part of the AnalyticsCreator design and used to generate the corresponding Power BI semantic-model artifacts.

AnalyticsCreator does not generate the report visuals themselves. Report pages, visuals, bookmarks and other report-specific elements remain part of Power BI development.

What AnalyticsCreator Generates

For Power BI development, AnalyticsCreator can generate:

Power BI Project output
Semantic-model definitions in native TMDL
Tables representing configured analytical structures
Relationships between model objects
Measures defined in AnalyticsCreator metadata
Hierarchies and related semantic-model metadata where configured

The generated semantic model provides the foundation for the Power BI solution. Power BI developers can then validate the model, configure the required connections and create the report layer in Power BI.

Power BI Projects, TMDL and PBIR

Microsoft Power BI Projects (PBIP) store Power BI development artifacts as source-control-friendly files and folders rather than only as a single PBIX file.

Within this structure, TMDL (Tabular Model Definition Language) represents the semantic-model definition. Power BI report definitions are maintained separately using Microsoft's report formats such as PBIR.

AnalyticsCreator generates the semantic-model side of this workflow. Report visuals and report-specific definitions are authored separately in Power BI.

Generation, Deployment and Execution

Generation, deployment and execution are separate activities.

AnalyticsCreator maintains the model definitions and generates the configured semantic-model artifacts.
The selected data platform stores and processes the underlying warehouse or analytical data.
Power BI consumes the data through the configured semantic model and provides report authoring and visualisation.
Git and CI/CD provide version control, review and controlled delivery of the generated definitions.

The exact warehouse architecture and Power BI connection model depend on the project design and selected Microsoft technologies. AnalyticsCreator does not require every project to use one fixed warehouse-layer topology.

Recommended Development Workflow

Define the solution in AnalyticsCreator. Import or maintain source metadata, mappings, business rules and the analytical model. Agree which definitions are owned by AnalyticsCreator.
Generate the required implementation artifacts. Generate the configured warehouse, loading and analytical outputs required by the project.
Generate the Power BI project. Use Create PowerBI project to generate the semantic model and measures in native TMDL.
Validate in Power BI. Open or integrate the generated project with Power BI Desktop, validate model connections, relationships, measures and other required model behaviour, and create the report visuals separately.
Review changes in Git. When AnalyticsCreator-owned definitions change, regenerate the model files and review the resulting Git differences before merging them.
Release through the configured Power BI or Fabric workflow. Deliver the reviewed definitions through the Git-connected workspace or configured CI/CD process and validate the resulting semantic model and reports.

Git and CI/CD

Because the generated Power BI semantic-model definitions are stored as text-based project files, they can be managed using normal Git development practices.

Teams can use this workflow to:

track semantic-model changes
work with branches and pull requests
review generated differences before release
maintain an auditable history of model changes
include the generated model definitions in configured CI/CD workflows

Git source control and workspace deployment remain separate steps. Committing a generated change does not by itself update a Power BI or Fabric workspace.

Regeneration and Ownership

AnalyticsCreator can regenerate the semantic-model definitions when the underlying AnalyticsCreator design changes.

Teams should therefore agree which definitions are owned by AnalyticsCreator and which are maintained manually in Power BI or other development tools. Generated files should be reviewed before release, particularly where manual extensions or downstream report dependencies exist.

Generation of TMDL files does not imply bidirectional synchronization of every external edit back into AnalyticsCreator.

Power BI Responsibilities

Power BI remains responsible for the reporting and operational parts of the solution, including the report layer and the Power BI-specific configuration required for the selected deployment.

Depending on the implementation, this can include:

report pages and visuals
report-specific configuration
connections and credentials
workspace configuration
refresh configuration
deployment through supported Microsoft mechanisms

Prerequisites and Considerations

Validate generated Power BI project output against the Power BI Desktop version and model features used by the project.
Keep generated and manually maintained definitions under an agreed ownership model.
Review regenerated files before merging or deploying changes.
Follow Microsoft's current PBIP, TMDL, Git integration and workspace requirements for the target environment.
Validate model behaviour and report dependencies after deployment.

Key Takeaway

AnalyticsCreator maintains the governed analytical design and generates the Power BI semantic model and measures as native TMDL. Power BI remains responsible for report authoring and the Microsoft-side runtime and workspace configuration, while Git and CI/CD provide the controlled path from generated model definitions to release.