Daten-WG

Francesco Bergamaschi joins from Italy to talk about DAX: the language at the core of every Power BI model, and the one most people feel they half-understand at best.





We get into:


- why calculated columns are just a tool and not a sin,


- the one rule that fixes most DAX performance problems,


- how to use ALLSELECTED without getting burned


- why CALCULATE beats variables for beginners,


- why DAX is "honest" and how it compares to SQL,


- whether DAX still matters in the age of AI


...and Francesco's teaching method – the five pillars of DAX and the path from a simple column to the measure magic.





Francesco Bergamaschi https://www.linkedin.com/in/francescobergamaschi/


Artur König


https://www.linkedin.com/in/datakoenigartur/





Meet Francesco and learn the five Pillars of DAX at the 14th and 15th October in Cologne:


https://www.daten-wg.com/details-registrierung/all-in-dax-pillars-en


https://www.daten-wg.com/daten-wg-2026








00:00 – Intro: an Italian DAX pro and why DAX tutorials all have an Italian accent


04:08 – Calculated columns: just another tool, not a sin


05:45 – Calculated tables and the danger of materializing whole tables


07:24 – Learning DAX in 2012, before Marco and Alberto's first book


09:03 – ALLSELECTED: how to use it without getting burned


11:03 – How do you teach DAX in a single day?


11:30 – The five pillars of DAX


13:19 – The teaching path: from columns to tables to the measure magic


21:10 – CALCULATE as the magic box – and why it beats variables for beginners


22:21 – "Variables are constant" and what makes DAX honest


24:56 – DAX vs SQL: three lines instead of twenty


25:26 – A language for the business user, not for storage


27:26 – The CALCULATE algorithm: a genuinely beautiful piece of engineering


30:13 – The Definitive Guide to ALLSELECTED – and getting a note added to it


32:01 – Is DAX still relevant in the age of AI? A new book from Marco and Alberto


33:56 – On the road: SQLBits, Data Point Prague, and a stop in Cologne


35:20 – Closing: paper, pencil, and good books








Links and sources


kubisco – Francesco's consultancy and DAX blog: https://www.kubisco.com/


The Definitive Guide to DAX, Third Edition (Marco Russo & Alberto Ferrari): https://www.sqlbi.com/books/the-definitive-guide-to-dax-third-edition/


The Definitive Guide to ALLSELECTED (SQLBI article): https://www.sqlbi.com/articles/the-definitive-guide-to-allselected/


SQLBI: https://www.sqlbi.com/


DAX Studio: https://daxstudio.org/


Data Point Prague 2026: https://datapointprague.cz/


SQLBits: https://sqlbits.com/

Daniel Marsh-Patrick joins from down under to talk about Deneb, the certified Power BI custom visual that turns Vega and Vega-Lite into a sandbox for almost any visual you can imagine.





We dig into declarative vs. imperative visualization, why Deneb is a side project he has been carrying for five years, what PBIR and the new JSON report format mean for visual developers, why Vega is unusually LLM-friendly, and where Deneb really shines: the engineering, environmental, and "non-business" charts that Power BI core visuals were never built for.





Guest: Daniel Marsh-Patrick https://www.linkedin.com/in/daniel-m-p/


Host: Artur König https://www.linkedin.com/in/datakoenigartur/





Daten-WG 2026 Conference, October 14-16: https://www.daten-wg.com/daten-wg-2026





00:00 – Intro: An international expert episode with Daniel from down under


02:55 – Daniel's path into Power BI visuals: Vega and Vega-Lite explained


05:13 – BI developer learning to be a web developer: where Power BI hits a wall


07:02 – Custom visuals, certification, and why proprietary visuals are tricky


09:09 – Why Deneb exists: certified custom visuals without the AppSource listing


10:49 – A side project for the community: five years of Deneb


12:59 – Declarative vs. imperative: why JSON beats coding every pixel


14:14 – PBIR and the new JSON report format for Power BI


17:32 – The big performance refactor and documenting Deneb's property structure


20:47 – Self-confessed non-coders build the most mind-blowing Deneb visuals


23:16 – Shoutouts to Kerry Kolosko and Ben Ferry


25:41 – Performance, interactivity, and the slow grind of community-driven roadmaps


26:44 – Vega and AI: why LLMs love declarative visuals


28:30 – Vega is everywhere: Looker Studio, Databricks, Microsoft research


30:30 – Why Deneb cannot have Copilot inside (and why it is not really a problem)


33:00 – Portability: copy a Vega spec from Databricks straight into Power BI


35:00 – When to reach for Deneb instead of stacking 28 measures for an IBCS bar chart


36:40 – Beyond sales and finance: engineering and environmental visuals


40:30 – Can I really maintain this? The third-party visual trade-off


42:00 – The world is broader than your next sales report


42:50 – Outro: Power BI & Fabric Summit and a shoutout to Reza Rad


44:00 – Difinity Conference and the open source spirit


45:11 – New Deneb release on the way with major performance improvements








Links and sources





Deneb documentation: https://deneb-viz.github.io/


Deneb on AppSource: https://marketplace.microsoft.com/en-us/product/power-bi-visuals/coacervolimited1596856650797.deneb


Vega: https://vega.github.io/vega/


Vega-Lite: https://vega.github.io/vega-lite/


Vega online editor: https://vega.github.io/editor/


Kerry Kolosko: https://kerrykolosko.com/


Ben Ferry – Power BI Ferry Tales: https://www.youtube.com/@powerbiferrytales


Microsoft Research – Data Formulator: https://www.microsoft.com/en-us/research/project/data-formulator/


IBCS: https://www.ibcs.com/


Power BI & Fabric Summit: https://globalpowerbisummit.com/

Mit Jasmin tauchen wir in ihr Herzensthema ein: Was uns Metadaten über Daten verraten: über Datenmodelle, über Skills im Team, über Reports, die niemand öffnet.





Wir reden über den Measure Killer, SharePoint-Listen als Mini-Datenkatalog, Datenqualität pro Use Case, Process Mining – und warum die meisten Unternehmen am ersten Schritt scheitern, weil sie zu lange über das richtige Tool diskutieren statt einfach loszulegen.





Gast: Jasmin Simader https://at.linkedin.com/in/jasmin-simader


Host: Artur König https://www.linkedin.com/in/datakoenigartur/





Daten-WG 2026 Konferenz am 14-16. Oktober - Agenda ist online:


https://www.daten-wg.com/daten-wg-2026





Jasmin ist bei einer Pre-Con dabei.


Mit dem Code SUPERJASMIN gibt es 20% Rabatt auf ihre Pre-Con





00:00 – Intro: Metadaten als Forensik – die fehlenden Puzzleteile


13:30 – Der Mini-Datenkatalog: SharePoint-Liste mit drei Spalten reicht für den Anfang


15:47 – Großkonzern oder Mittelstand: Wer profitiert mehr von Metadaten?


20:14 – Erste Schritte: MeasureKiller dranhängen und Status erheben


20:37 – Im BI-Team fragen, was Metadaten überhaupt sind


21:03 – Outro: Inventarisierung, Bucketlist und Schulterblick zu Benni De Jagere





Power BI Vibes: https://www.youtube.com/@BIVibes-JasminSimader


Measure Killer: https://en.brunner.bi/measurekiller


Benni De Jagere: https://bennidejagere.com/


Microsoft Fabric Metadata Scanning: https://learn.microsoft.com/en-us/fabric/governance/metadata-scanning-overview

Die erste FabCon 2026 ist durch – und Gabi Münster ist wieder im Quarterly, um mit Artur die Ankündigungen einzuordnen. Von Custom Totals in Power BI über Translytical Task Flows GA bis hin zu OneLake Security und Fabric Planning mit Lumel.





Und zwischendurch geht es darum, was das alles für Pro-User, für Dataflow-Strecken und für die Frage "wie offen ist Fabric wirklich?" bedeutet – inklusive Migration Accelerator, SCD Typ 2 im Copy Job und Auto-Scaling als Quality-of-Life-Feature.





Gast: Gabi Münster https://www.linkedin.com/in/gabimuenster/


Host: Artur König https://www.linkedin.com/in/datakoenigartur/





Daten-WG 2026 Konferenz am 14-16. Oktober: https://www.daten-wg.com/daten-wg-2026





Dataflow Migrations-Assistent in der Fabric Toolbox:


https://github.com/microsoft/fabric-toolbox/tree/main/accelerators/DFG2-migration-accelerator





00:00 – Intro: FabCon 2026 und die erste Welle der Ankündigungen


00:31 – Power BI: Custom Totals in Tabelle und Matrix


03:07 – Visual Calculations als Shortcut, nicht als Modell-Ersatz


03:57 – März-Update: Neues UI-Default und grauer Hintergrund


04:25 – PBR als neuer Standard für Power BI-Modelle


05:55 – Pro-Developer-Format: Warum strukturierter Code jetzt überall zählt


07:53 – Translytical Task Flows sind GA


09:52 – Lego-Roboter per Translytical: Premiere beim Data Grillen


11:00 – Dataflows Gen 1 wird Legacy – was heißt das wirklich?


15:37 – Fabric Toolbox: Migration Accelerator für Gen 1


16:48 – Was Pro-User jetzt im Blick haben müssen


21:26 – SCD Typ 2 im Copy Job (Preview) – das Ende vieler Notebooks


24:38 – Plattform: SAP GA, Snowflake und Databricks bidirektional, Airflow und DBT


26:53 – OneLake Security: Interoperabilität statt Vendor Lock-in


30:53 – Was in OneLake Security noch fehlt – und warum GA trotzdem zählt


33:30 – Quality of Life: Papierkorb und Overage Billing


42:00 – Fabric Planning mit LUMEL: First-Party-Workload ohne Extra-Lizenz


53:23 – Planning vs. Translytical Task Flows: zwei Werkzeuge, zwei Jobs


54:25 – Gabis FabCon-Highlights: OneLake Security und Governance


56:00 – Daten-WG 2026, Data Grillen und SQL Bits