Гаспър Каменшек Мастърклас

Modern DAX A–Z: Write Better Measures, Build Smarter Models

A one-day masterclass focused on the modern Power BI Desktop and DAX experience. The course highlights the most important Power BI Desktop and DAX improvements from the last two years, including DAX Query View, calculation groups, visual calculations, recent DAX capabilities, and the shift toward AI-ready semantic models.

Basic topics list

  1. What’s New in Modern Power BI Desktop and DAX
  2. Designing AI-Ready Semantic Models
  3. Writing Better Measures
  4. DAX Query View for Testing and Debugging
  5. Calculation Groups for Reusable Business Logic
  6. Recent DAX Functions and Modern Calculation Patterns
  7. Visual Calculations: Where They Fit
  8. Putting It All Together: A Modern Power BI Development Workflow

Detailed topics with descriptions

Гаспър Каменшек Мастърклас

Part 1: What’s New in Modern Power BI Desktop and DAX

Description
This opening section sets the context for the day. Power BI Desktop has changed significantly, and modern DAX work is no longer only about writing measures in the formula bar and checking the result in a visual. The newer experience encourages testing, reusable logic, clearer semantic models, and more deliberate separation between model-level calculations and visual-level calculations.

Topics

  • What “modern DAX” means today
  • The shift from report-building to semantic model design
  • Why newer Power BI Desktop features matter for everyday BI development
  • DAX Query View as a better way to test and understand calculations
  • Calculation groups as a way to centralize repeated business logic
  • Visual calculations as a new visual-level calculation layer
  • Recent DAX capabilities and where they are practically useful
  • The growing importance of semantic models for Copilot and AI-driven querying

Key takeaway

Attendees should understand that modern Power BI work is not just about writing individual measures. It is about building a reliable semantic layer that supports reports, users, and AI-assisted analysis.

Part 2: Designing AI-Ready Semantic Models

Description
This is the modeling foundation of the masterclass. The course should emphasize that AI querying does not magically fix a messy model. If tables, columns, relationships, measures, and descriptions are unclear, AI experiences will struggle to return reliable answers. A strong semantic model needs clear business meaning, not just technically correct relationships.

Topics

  • What makes a model “semantic”
  • Star schema as the foundation for reliable DAX
  • Clear fact and dimension table design
  • Business-friendly naming conventions
  • Hiding technical columns from report users
  • Creating measures that represent business concepts
  • Using descriptions and metadata to explain fields and measures
  • Designing models that users can query naturally
  • Why AI experiences depend on clear relationships, measures, names, and definitions
  • Common modeling mistakes that make both DAX and AI querying harder

Practical examples

  • Renaming source-system fields into business-friendly names
  • Turning raw columns into curated measures
  • Hiding unnecessary technical fields
  • Creating a clean measure table or measure organization structure
  • Adding descriptions to important business measures

Key takeaway

A smart semantic model is not just a data structure. It is a communication layer between the data, report developers, business users, and AI experiences. Microsoft’s Copilot guidance makes the semantic model central to answering data questions, which makes model clarity and metadata more important.

Part 3: Writing Better Measures

Description
This section connects the course title directly to day-to-day DAX work. The focus should be on writing measures that are readable, reusable, testable, and aligned with business logic. This should not become a beginner DAX syntax session, but it should refresh the core patterns that matter most in modern models.

Topics

  • Measures as the primary place for business calculations
  • Base measures vs derived measures
  • Measure branching
  • Using VAR for readability and debugging
  • Naming measures clearly
  • Organizing measures into folders or logical groups
  • Avoiding repeated logic across many measures
  • Writing measures that are easy to test in DAX Query View
  • Choosing whether logic belongs in Power Query, the model, a measure, a calculation group, or a visual calculation

Practical examples

  • Base measure: Revenue
  • Derived measure: Revenue YTD
  • KPI measure: Actual vs Plan
  • Ratio measure: Over/Under Plan %
  • Dynamic text measure: Report subtitle or selected-period label

Key takeaway

Better measures are not just shorter formulas. They are measures with clearer intent, less duplication, better structure, and more predictable behavior.

Part 4: DAX Query View for Testing and Debugging

Description
This should be one of the signature sections of the day. DAX Query View gives Power BI developers a more professional way to test, inspect, and validate calculations directly inside Power BI Desktop. It moves DAX development away from “put it in a visual and hope it works” toward a more deliberate testing workflow. Microsoft describes DAX Query View as the Power BI Desktop experience for writing and running DAX queries against the semantic model.

Topics

  • What DAX Query View is
  • Why it changes the DAX authoring workflow
  • Testing measures outside visuals
  • Understanding query results with EVALUATE
  • Using SUMMARIZECOLUMNS for grouped results
  • Inspecting filter context through queries
  • Comparing expected vs actual measure results
  • Using queries to validate totals, subtotals, and slicer behavior
  • Saving useful DAX query tabs in the Power BI file
  • Using DAX Query View as a learning, debugging, and documentation tool

Practical examples

  • Test a measure by product, customer, or date
  • Validate a percentage-of-total measure
  • Inspect why a total behaves differently from row-level values
  • Compare two versions of a measure
  • Use DAX Query View to explain filter context

Key takeaway

DAX Query View should become part of the regular development workflow for serious Power BI authors.

Part 5: Calculation Groups for Reusable Business Logic

Description
Calculation groups deserve a central place in the course. They are one of the strongest examples of “write better measures, build smarter models” because they reduce repeated measures and centralize logic. Native Power BI support also makes them more approachable for users who previously relied on external tools. Microsoft documents creating calculation groups in Power BI, including calculation items, selected measure behavior, and dynamic format strings.

Topics

  • What calculation groups are
  • Why they reduce duplicated measures
  • Calculation items and selected measures
  • Time intelligence calculation groups
  • Actual, Budget, Forecast, and Variance patterns
  • Dynamic format strings
  • Precedence and calculation group behavior
  • When calculation groups are a good idea
  • When calculation groups may add too much complexity
  • How calculation groups support cleaner, more scalable semantic models

Practical examples

  • One calculation group for time intelligence:
    • Current
    • Previous period
    • Year-to-date
    • Year-over-year
    • Year-over-year %
  • One calculation group for scenarios:
    • Actual
    • Plan
    • Variance
    • Variance %
  • Dynamic format string example:
    • Currency measures
    • Percentage measures
    • Whole-number measures

Key takeaway

Calculation groups help move repeated logic out of individual measures and into a reusable calculation layer.

Part 6: Recent DAX Functions and Modern Calculation Patterns

Description
This section should highlight recent and important DAX capabilities without becoming a function-by-function reference. The emphasis should be on what these functions make easier, where they fit, and how they change practical modeling or reporting patterns. Microsoft maintains a dedicated list of new and updated DAX functions, and notes that new DAX functionality is typically introduced first in Power BI Desktop.

Topics

  • Why new functions matter, but not every function needs deep coverage
  • Window-style DAX thinking
  • Practical use cases for functions such as:
    • OFFSET
    • INDEX
    • WINDOW
    • ORDERBY
    • PARTITIONBY
    • MATCHBY
    • RANK
    • ROWNUMBER
  • Comparing rows within a partition
  • Ranking and ordering patterns
  • Period-over-period style calculations
  • When newer functions simplify older, more complex DAX patterns
  • Awareness of DAX user-defined functions for reusable parameterized logic

Practical examples

  • Rank products within category
  • Compare current row with previous row
  • Calculate moving or rolling windows
  • Retrieve a value from a neighboring period
  • Create more readable ranking logic

Key takeaway

The focus is not memorizing new functions. The focus is recognizing when recent DAX capabilities make business calculations clearer, shorter, or easier to maintain.

Part 7: Visual Calculations: Where They Fit

Description
Visual calculations are important to include because they are one of the clearest recent changes in how Power BI users can write DAX. But they need careful positioning. They are useful for visual-specific logic, but they should not replace properly modeled business measures. Microsoft describes visual calculations as DAX calculations defined and run directly on a visual rather than stored in the semantic model.

Topics

  • What visual calculations are
  • How they differ from model measures
  • Why they are calculated at the visual level
  • Where visual calculations are useful
  • Where visual calculations are risky
  • How visual calculations relate to reusable business logic
  • Visual-level running totals, rankings, and comparisons
  • Why governed KPIs should usually remain model measures
  • How to decide between a measure, calculation group, and visual calculation

Practical examples

  • Running total in a specific visual
  • Visual-level rank
  • Difference from previous item in a matrix
  • Percent of visual total
  • Quick exploratory calculation that does not need to become a governed measure

Key takeaway

Use visual calculations for visual-specific analysis. Use model measures and calculation groups for reusable, governed business logic.

Part 8: Putting It All Together: A Modern Power BI Development Workflow

Description
This final section turns the day into a repeatable framework. It should help attendees leave with a clear way to apply what they learned in their own models.

Topics

  • Start with business questions, not visuals
  • Design a semantic model around business concepts
  • Build clean base measures
  • Use measure branching for derived logic
  • Use calculation groups for repeated calculation patterns
  • Test and validate calculations in DAX Query View
  • Use newer DAX functions where they simplify the solution
  • Use visual calculations only for visual-specific logic
  • Add names, descriptions, folders, and metadata to support users and AI querying
  • Review the model for clarity, reuse, performance, and maintainability

Suggested final framework

  1. Model clearly
    • Use a clean star schema
    • Name tables and fields in business language
    • Hide technical clutter
    • Add descriptions where meaning is not obvious
  2. Define logic once
    • Create reusable base measures
    • Branch into derived measures
    • Use calculation groups for repeated patterns
    • Avoid copy-paste measure sprawl
  3. Validate before publishing
    • Test measures in DAX Query View
    • Check totals and filter behavior
    • Confirm visual calculations are used only where appropriate
    • Make the final semantic model understandable for users, developers, and AI

Software Requirements

  • Power BI Desktop (most current version available)
  • Windows 11 (64-bit), or Windows 10 (64-bit, version 21H2 or newer)

Hardware Requirments

  • PC laptop, minimum 16 GB RAM recommended;

! Mac versions are not supported, and Power BI Desktop does not have a Mac version.