Next week I am in Mechelen for dataMinds Connect 2026, 12 to 14 October. On Tuesday 13 October at 16:10 I present “Personalized Semantics with User-Context-Aware calculated columns” in the Safety Institute room (Verloren Zoon). If you are there, come and say hello.
What the session covers
The talk is about calculated columns that do not resolve to the same value for every user. Power BI now exposes an Expression Context property on calculated columns. With it, functions such as USERCULTURE(), USERPRINCIPALNAME(), USERNAME(), USEROBJECTID() and CUSTOMDATA() can live in a column, not only in a measure.
Translation is the obvious case: labels and captions that follow the person reading the report. Personalisation goes further: logic that depends on who is looking, without a separate semantic model for each audience. The same pattern works across Direct Lake on OneLake, Import and DirectQuery, but the moment the column is evaluated differs by mode. We will look at materialised versus unmaterialised behaviour, and what that does to refresh and query performance.
It is a level 300 session. The demos assume you already model in Power BI. I will not start from a blank canvas.
For a business reader, the point is simpler. One semantic model can show each person labels and logic in their own language and context. No copy per audience, nothing to keep in sync.
Why dataMinds
dataMinds Connect is a volunteer-run Microsoft Data Platform conference. Three days in Mechelen: deep dives on Monday, conference sessions on Tuesday and Wednesday. It is a room full of people who build this for a living, which is why I wanted to be there.
Come and talk about where BI is going
I am there all three days. Happy to talk Power BI, Microsoft Fabric, semantic models and Direct Lake, and where BI, data and analytics are heading. Bring a modelling question or an opinion about the future of BI. I have a few of my own.
Find me after the session or on LinkedIn, or get in touch if you would rather plan a slot.
