An Analysis Model is the semantic layer between a data source and Datafor reports or AI-assisted analysis. It turns database tables and columns into reusable business objects with consistent joins, names, calculations, formats, and definitions.

What a model contains

Object	Purpose
Table	A source table or database view, or a model-defined SQL View.
Relationship	Defines how two tables are joined and the cardinality of each side.
Dimension	Groups descriptive attributes used to filter, group, and drill into data.
Attribute	A business-facing field such as Store, Product, Status, or Order Date.
Hierarchy	Orders attributes into drill paths such as Year → Month → Day.
Measure	Aggregates a source field with a method such as Sum, Count, or Distinct Count.
Calculated measure	Evaluates a reusable model-level expression at query time and can reference Measures.
Business semantics	Adds descriptions, aliases, roles, units, direction, and recommended dimensions.
Enterprise metric binding	Connects a model measure to a governed definition in Metrics Library.

Modeler workspace

Area	Use it to
Toolbar	Refresh connection metadata, undo or redo model edits, open Data security setting (gear), create a SQL View, and save the model.
Data panel	Select a datasource and schema, search tables, preview data, and add tables to the model.
Canvas	Inspect tables, arrange the model, and create or edit relationships.
Analysis model	Manage Dimensions, Attributes, Hierarchies, Measures, and calculated measures.
Attributes	Edit the selected object's core properties, business semantics, advanced settings, and metric governance.
Bottom bar of the canvas	Model diagnostics and Metric bindings tabs: review diagnostics and enterprise metric bindings.

Recommended modeling sequence

1. Add only the tables required for the intended analysis.
2. Define and verify every relationship.
3. Remove semantic objects that do not match each table's role, manually or with a reviewed Tidy preview.
4. Build useful Dimensions, Attributes, Hierarchies, and Measures.
5. Configure time roles and add business descriptions and other semantic metadata.
6. Resolve model diagnostics.
7. Save the model and validate representative queries or reports.

Diagnostics detect many structural and semantic problems, but they cannot prove that a join or business definition is correct. Validate important totals against a trusted source before publishing the model.

Related topics

• Managing Analysis Models
• Creating an Analysis Model
• Working with Tables and the Canvas
• Creating SQL Views
• Establishing Table Relationships
• Advanced Relationship Modeling
• Tidy Model into a Star Schema
• Creating Hierarchies
• Time Semantics and Default Time Settings
• Calculated Columns
• Measures and Calculated Measures
• Business Semantics for AI
• Model Diagnostics