Power BI

This page describes how AnalyticsCreator generates and integrates analytical models for Power BI.

Overview

AnalyticsCreator supports Power BI as a target for analytical consumption by generating semantic models based on the data warehouse structure. These models define dimensions, measures, and relationships that can be used directly in reporting.

AnalyticsCreator does not execute data processing inside Power BI. Instead, it generates semantic structures that are consumed by Power BI after data has been processed in the underlying data warehouse platform.

Supported Services and Components

  • Power BI semantic models (tabular models)
  • Power BI datasets
  • Power BI Desktop and Power BI Service
  • DirectQuery and import modes

What AnalyticsCreator Generates

For Power BI, AnalyticsCreator generates:

  • Semantic models:
    • Dimensions and facts mapped from DM layer
    • Relationships between entities
    • Measures and calculated fields
  • Model structure:
    • Hierarchies (e.g. date hierarchies)
    • Attribute groupings
    • Metadata descriptions
  • Deployment-ready artifacts:
    • Tabular model definitions
    • Integration with deployment workflows

Supported Modeling Approaches

  • Dimensional modeling (star schema)
  • Measures derived from fact tables
  • Hierarchical dimensions (e.g. time, geography)

Data Vault models are not directly exposed to Power BI. Instead, dimensional structures are generated from CORE layers and used as the basis for semantic models.

Deployment and Execution Model

AnalyticsCreator separates semantic model generation from data processing:

  • AnalyticsCreator generates the semantic model definition
  • The model is deployed to Power BI or an analytical engine
  • Data processing happens in the underlying data warehouse
  • Power BI consumes the processed data through the semantic model

Execution flow:

  • Data is processed in STG, CORE, and DM layers
  • Semantic model references DM structures
  • Power BI queries the model or underlying data source

CI/CD and Version Control

  • Semantic models are generated from metadata definitions
  • Model definitions can be included in deployment pipelines
  • Integration with version-controlled AnalyticsCreator projects

Connectors, Sources, and Exports

Data sources for Power BI

  • SQL Server
  • Azure SQL
  • Microsoft Fabric
  • Other supported warehouse targets

Export targets

  • Power BI semantic models
  • Tabular model deployments

Prerequisites, Limitations, and Notes

  • Power BI requires access to the deployed data warehouse
  • Model performance depends on underlying data model design
  • DirectQuery performance depends on source system performance

Design considerations:

  • Use DM layer as the source for semantic models
  • Avoid exposing STG or CORE layers directly
  • Ensure measures are defined consistently

Example Use Cases

  • Generating a Power BI-ready star schema from a data warehouse
  • Automating creation of measures and relationships
  • Standardizing reporting models across projects
  • Reducing manual modeling effort in Power BI

Platform-specific FAQ

Does AnalyticsCreator replace Power BI modeling?

AnalyticsCreator generates the base semantic model, including dimensions, relationships, and measures. Additional report-specific logic can still be implemented in Power BI if required.

Where are measures defined?

Measures can be generated as part of the semantic model based on metadata definitions and can be extended in Power BI.

Can Power BI connect directly to CORE or STG?

This is technically possible but not recommended. The DM layer should be used as the primary source for reporting.

Proof Assets

  • Generated semantic models include dimensions, measures, and relationships based on metadata
  • Demo scenarios show end-to-end flow from source to Power BI-ready model

Related Content

Commercial Solution Page

For product-level positioning and overview, see the AnalyticsCreator Power BI solution page.

Key Takeaway

AnalyticsCreator generates semantic models for Power BI based on data warehouse structures, while data processing remains in the underlying platform.