powerbiweekly.info

#352 - 7th April 2026

2026-04-07T00:00:00.00Z

No official blogs this week, but you'll notice that there's quite a bit more content than usual. That's because I've been on the hunt for some additional feeds and found a handful of new YouTube channels publishing great content. Plenty of end-to-end demos, AI-assisted Power BI development and general tips and tricks to complement the existing content. As ever, if you know of any creators that I'm not currently picking up, then please do share them with me so we can share more widely!

As for the highlighted content this week: Kurt Buhler has seemingly been busy working on the pbir-cli, a CLI tool (with integration with AI assistants) that helps with many operations, particularly those mundane and time-consuming reporting tasks. Definitely worth checking out! On YouTube, Okeh Efasa has shared the video No More Boring Tables! Power BI Conditional Formatting (Power BI DAX + Claude AI), demonstrating a great (and simple) us-case for AI assistance in your reports by making table visuals much more interesting with a few smart SVG-outputting measures. Also on YouTube, and a bit of a fun one, Ned Charles has demonstrated How to build a Power BI Semantic Model on an iPad (that's right, an iPad) to put the new online semantic-modelling tooling through its paces. Ned's conclusion probably won't surprise you, but it was an interesting experiment nonetheless! Finally, Chris Webb has revisited the topic of Role-Playing Dimensions In Fabric Direct Lake Semantic Models, showing some TMDL tricks to achieve the target modelling behaviour necessary to use the same Lakehouse table for two Power BI tables.

#351 - 31st March 2026

2026-03-31T00:00:00.00Z

A handful of official blogs this week. On the Fabric side (but relevant to Power BI) is the release of Fabric CLI v1.5, including full support for all Power BI REST API operations (and a couple of new capabilities thrown in too!) We've also had a deep-dive article describing how to Chat with Copilot inside a report on the Power BI mobile app (Preview) (announced at FabCon), and we've had an article around Dataflows: Thank you for eight years of Gen1—and why Gen2 is the future, suggesting that a retirement date for Gen1 is imminent, but they will be supported for the "foreseeable future".

Outside of Microsoft this week, Marco Russo has succinctly described the differences between DAX user-defined functions (UDF) vs. calculation groups, Tharun Kumar Ravikrindhi has told us to Stop Writing Format Strings for Every Measure | Use This One DAX UDF Function Instead, and Dániel Patkós has continued his series on Power BI CI/CD in GitHub Actions with the article: Making BPA Results Actionable with Automated PR Comments, a great way to enforce quality gates proactively in a discoverable manner!

#350 - 24th March 2026

2026-03-24T00:00:00.00Z

This week we have all the FabCon related Power BI content, in case you missed anything last week. The main blog, written by Arun Ulag, can be found here: FabCon and SQLCon 2026: Unifying databases and Fabric on a single data platform, putting databases right at the forefront of this year's messaging. You won't see much Power BI content in there, but if you're using Fabric (which I'm sure many of you are), it's well worth taking a read of that to get your summary. If you want the Power BI round-up, then that's been written by Mohammad Ali, titled: Semantic Layers: The foundation of enterprise AI, highlighting the main announcements:

The GA of Translytical Task Flows
The desperately-needed "Modern visual defaults" (see a deep dive here: Deep Dive into Modern Visual Defaults and Customizing Theme Improvements (Preview)). By default, reports now look much slicker and aesthetically pleasing (IMHO)
The preview of Report Copilot for Mobile
The introduction of TMDL View on the Web (Preview), bringing the power of scripted changes to your model into the browser, where Version History is supported and previewing changes is supported through the "View" mode
The GA Direct Lake on OneLake
The Preview of Direct Lake Calculated Columns, including a new "user-context aware" capability which allows for "dynamic" columns based on user-specific DAX functions like UserCulture(), UserPrincipalName(), and CustomData() 🤯
The GA of various Table Visual Updates (including new, easy options for overriding totals)

Even more updates can be found in the Power BI March 2026 Feature Summary, which was also released last week. Another announcement over on the Fabric blog was the GA of the On-premises data gateway auto-update (admin triggered), enabling on-demand manual or programmatic updates to your Gateway. And today we've been told about the GA of Workspace Tags, providing even more options for organising and describing your workspaces.

Finally, a couple of non-FabCon bits of content this week. Just Blindbæk has written an interesting blog about how to use and abuse the Semantic Model Scale-out feature to Reduce Refresh Memory on Large semantic model in Microsoft Fabric, by effectively clearing the model's data before re-hydrating it with your data. Elsewhere, Ben Gribaudo has shown a way to achieve Less GUIDs: Relative Current Workspace & Items By Name. If anyone has used a Dataflow Gen2 or Power Query to retrieve items from a Lakehouse, you'll benefit from Ben's trick highlighted in that article!

If you're interested in all things Microsoft Fabric, don't forget to sign up for our new newsletter Fabric Weekly. We'll be moving all Fabric content over from Azure Weekly in the near future. Don't worry - Power BI Weekly is remaining as-is, for the Power BI purists out there!

#349 - 17th March 2026

2026-03-17T00:00:00.00Z

Not a FabCon special this week - that will come mid next week with the full round up of all Power BI related announcements. This week, just a handful of community content to whet your appetite. First of all, Jiří Neoral has written an interesting article (in Czech. but use translate if necessary!) about using standard Gen AI chat experiences to Generate a Power BI layout, comparing the likes of Claude, Gemini, ChatGPT and Copilot as to how well they generate a report page using the same prompt across each. I think this is an interesting experiment, given that it's unlikely business users will get familiar with the MCP tools that would be better suited for these types of operations, so (IMHO) comparing the standard chat experiences is a valid test.

Elsewhere this week, Zoe Douglas has written about MoM, MoM%, YoY, and YoY% DAX measure approaches (providing a handy TMDL snippet to generate all at once), Bas Dohmen has shared a means by which you can use a DAX Measure to Send Emails Directly from Power BI (or, rather, "draft" emails), and Gary Carpenter has demonstrated how to Build a Liquid Glass Navigation Menu in Power BI.

#348 - 10th March 2026

2026-03-10T00:00:00.00Z

A short edition this week with no announcements to share. So let's highlight a few articles. First of all, Teo Lachev has provided an overview of Direct Lake Composite Models (currently in preview), comparing and contrasting them to Direct Lake-only models and proposing scenarios where you might like to consider the composite option. Elsewhere this week, a couple of Field Parameters videos: Reid Havens starts with a good overview of Power BI Field Parameters vs Calculation Groups (while showing off some new web-dev skills!) and Nadim Abou-Khalil has shared a Field Parameters Step-by-Step - Dynamic Reports Made Easy, showing how you can create arbitrary groupings over field parameters to achieve bulk showing/hiding of fields on your visuals. Nice trick!

#347 - 3rd March 2026

2026-03-03T00:00:00.00Z

This week we've received the Power BI February 2026 Feature Summary, including a few interesting reporting features such as GA of Filter or enter data in reports with Input slicer (formerly the Text slicer) and the ability to Paste selections into any report slicer, opening up this capability to the standard slicer. A bit more detail on the input slicer is available in the article Input slicer: Filter reports and collect user input.

Outside the feature summary this week, there's been the announcement about the ability to Evaluate Power Query Programmatically in Microsoft Fabric (Preview) whereby the Fabric team has exposed an API on top of a Dataflow to allow you to execute arbitrary queries using the Dataflow engine. Opens up a lot of automation scenarios if you're familiar with Power Query! Elsewhere this week, James Serra has written a useful blog comparing Building Power BI Reports: Desktop vs Fabric (given that the feature sets are somewhat converging) and Marco Russo has written about a simple way to troubleshoot your tricky measures, writing about Debugging DAX variables using TOJSON and TOCSV (a trick I still find useful since you don't need to try to recreate your evaluation context elsewhere!)

#346 - 24th February 2026

2026-02-24T00:00:00.00Z

Another small edition this week, so I'll only highlight a couple of articles. First of all, Marc Lelijveld has shown how you can use the Power BI MCP server along with GitHub Copilot to Generate ERD diagrams for your Power BI Semantic Model, whereby a draw.io diagram is created which can be edited and easily reused across projects. The interesting thing here is that, at least initially, the tool assumes you already have a Power BI model. But what about the Entity Relationship Diagrams created during design phase (before even your reporting lakehouse/database has been created)? What tools do you all use to visualize and document your data models ahead of time? The beauty of GenAI means that even without the MCP Server that interacts with your semantic model, you could still use GitHub Copilot (or other GenAI tool) to generate an ERD, provided you've given it enough context about your semantic model (such as a data dictionary). But how do you create them? Do you use Mermaid diagrams? Visio? draw.io? Miro/Mural? Would love to hear what tools you prefer!

The second article I'd like to highlight is another by Jen Stirrup, who has written about style over substance in the blog Is Your Power BI a Spotlight or an Instagram Filter? Jen stresses the importance of not "sweeping issues under the rug", suggesting that identifying and addressing flaws in your E2E data pipeline is a necessary and natural part of a report's lifecycle, and that "touching up" reports often causes fundamental issues in your report's data model/visualizations, or worse, confusion in the end-user's interpretation of your reports. Worth a read!

#345 - 17th February 2026

2026-02-17T00:00:00.00Z

A pretty quiet week this week. Only one official blog which comes from the Fabric side-of-things, which is a detailed walkthrough of how to Enrich Power BI reports with machine learning in Microsoft Fabric. In that post, Ruixin shows how you can connect to data in your semantic model using Semantic Link, perform data exploration using notebooks, create a model using Fabric ML and native MLflow experience, enable real-time scoring and feed these results back into another semantic model. Interesting stuff! On the community front, Mike Carlo and Tommy Puglia have spoken about Central BI & Workspace Strategies in Ep. 501 of the Explicit Measures podcast, a topic they've discussed before but one worthwhile revisiting with the additional architectures introduced by Microsoft Fabric. Moving on - those of you with a keen eye will have noticed the Fabric Workspace Settings Update: "License Type" Renamed to "Workspace Type" highlighted here by Nicky van Vroenhoven, whereby the Fabric/Power BI teams endeavour to disambiguate licensing concerns with the type of workspace you create with some simple terminology and dialogue design updates. I like this seemingly-trivial update, and hope that it reduces confusion for teams in the future. Finally this week, Chris Webb has gone into detail about Measuring Power BI Report Page Load Times, providing a Power Query script to plug into your Performance Analyzer exports in order to gain a more representative understanding on how long report interactions take for your end users.

#344 - 10th February 2026

2026-02-10T00:00:00.00Z

In the past week we've had a few official updates, the most exciting of which being the General Availability of Semantic Link, which is a very welcome announcement! Semantic Link vastly simplifies your automation scenarios while interacting with various Fabric artifacts, and now it's fully supported and available on Spark Runtime 2.0. Sadly it's still only a subset of functionality supported in Semantic Link Labs, but naturally that library can move at a pace of innovation that is hard to mirror for a fully supported service in an Enterprise platform like Fabric. Still, great news! Elsewhere this week, we've been told about the Deprecation of old Excel and CSV import experience in Power BI Service and the new ability to perform a Manual update for on-premises data gateway (Preview), which is the first step towards a fully automated update experience for your Gateways. A sigh of relief for many admins, I should think!

On the community front this week, I'd like to highlight a few articles from across the Power BI topic-spectrum. Firstly, Jen Stirrup has written about the recently-revealed phishing attacks utilising the Power BI Subscription feature, of which I was not aware. Jen ties this back into your Governance practices (both business processes and technical processes) to give some great tips as to how to guard against these types of attacks as best as possible. Onto some more blogs: Ben Gribaudo is back enlightening me of more Power Query (M) features I wasn't aware of: Table Interceptors. These are particularly helpful when overriding transformations that can't be natively folded on your source, but you can implement a custom handler that can do just that. Finally, Nicky van Vroenhoven has written a informative blog on Transitioning to the New Power BI Enhanced Report Format (PBIR) - What You Need to Know, particularly worth reading if you're a tenant admin and not entirely sure of the consequences of doing so.

#343 - 3rd February 2026

2026-02-03T00:00:00.00Z

On the official blog, Zoe Douglas has provided a Deep dive into composite semantic models with Direct Lake and import tables - the new public preview capability allowing you to mix Direct Lake and Import tables and achieve the expected performance with the creation of regular (rather than limited) relationships. Marco Russo has shown us how to Show transaction details on the matrix visual in Power BI, stepping through various DAX calculations culminating in the optimal method. Just Blindbæk has written about Bringing Databricks Metric Views into the Power BI world with the new Tabular Editor 3 (Commercial) feature called "Semantic Bridge", which offers automatic conversion of the two data model representations. Finally, the Icon Map team has written an interesting article about Gradient colour schemes in Power BI: keep your map and legend colours aligned, describing where gradient legends can fall short in your dashboards and how instead to use them effectively.

#342 - 27th January 2026

2026-01-27T00:00:00.00Z

This week we've been given the first Power BI January 2026 Feature Summary, though it's quite thin on the ground. The two main updates regard the new Markers in Azure Maps for Power BI (allowing you to use arbitrary icons/images for your mapped-out categories), and the GA of Modern Visual Tooltips in Power BI. There's also been the Power BI Report Server January 2026 Feature Summary introducing a bunch of visualisation and reporting enhancements based off recently released Power BI Desktop functionality, and also the On-premises data gateway January 2026 release.

As for external articles this week, I'd like to highlight a great article by Paul Turley, who's written about Power BI Fabric Patterns: Guidance for the Visual [Design] Impaired, discussing the various elements that need to be considered when designing an impactful, comprehensible and accessible report. Finally, Chris Webb is back making us aware of another gotcha, in his blog: Calculate(), DAX Fusion And Filters On 0 In Power BI, highlighting the importance of using strict equalities if you're filtering by integer values in your DAX measures.

#341 - 20th January 2026

2026-01-20T00:00:00.00Z

Quite a short edition this week with no official announcements. So, straight onto the set of highlighted articles. Gary Carpenter has released the final video in his four part Power BI Card series showing us how to Build a Pro-Level Power BI Card from Scratch, Karin Szilágyi details the importance of good alt text when designing Power BI reports for accessibility with her blog Don’t Make My Mistakes: A Key Lesson on Alt Texts for Power BI & DataViz Contests. On the Power Query side, Ben Gribaudo has written an article on Power Query's column types: Column Types Don’t Matter, or Do They? discussing why column types exist in Power Query and when they become important. Finally, the Learning Science channel walks through how to Copy a semantic model in one go from one Power BI report to another using TMDL.

If you're interested in all things Microsoft Fabric - don't forget to sign up for our new newsletter - Fabric Weekly - which we'll start publishing in the next month or so.

#340 - 13th January 2026

2026-01-13T00:00:00.00Z

No official announcements this week, other than the fact that the Power BI Dataviz World Champs is back with a kick-off session on Thursday, if you're interested! So straight onto the highlighted content this week. On the visual front, Gary Carpenter has continued his very useful Power BI Card series with Part 3: Advanced Card Layouts, Design, and Hero Images, and Bas Dohmen has shown us a Power BI Progress Bar You Haven't Seen Before entirely using native visuals with some formatting tricks (as per usual!) On the Power Query side, Ahmed Essam has written part 2 of the Power Query series, sharing Bulletproof Power Query (Part 2): A Smart, Fuzzy-Match Rename Function to gracefully handle source-column name changes. And finally, Luca Liu has provided a DAX function to Calculate a Dynamic Truncated Mean in Power BI Using DAX - an alternative to the arithmetic mean which removes extreme outliers from the calculation.

#339 - 6th January 2026

2026-01-06T00:00:00.00Z

Welcome to the 339th edition of Power BI Weekly! I hope everyone had a restful break and you're all set for 2026. There haven't really been any official bits of Power BI content since the last edition, other than the On-premises data gateway December 2025 release, which highlights a new manual "update" option that is shown in both the Gateway UI and accessible via the API. The highlighted content this week is quite varied: first of all, Ben Gribaudo is back with an M deep-dive, this time enlightening us on all the intricacies of Type Equality, Andrzej Leszkiewicz has walked us through How to install DAX Lib packages using Power BI Desktop or Tabular Editor or AI, Marco and Alberto have shown us the importance of understanding Many-Many relationships in the article Analyzing the performance of limited and regular relationships, and Gary Carpenter has shared with us the first couple of episodes in a 4-part series on how to Level Up Your Power BI Cards.

#338 - 16th December 2025

2025-12-16T00:00:00.00Z

Welcome to the 338th edition of Power BI Weekly! This will be the last edition of 2025 as I head on my holidays for the next couple of weeks. As ever, thank you for another year of your readership, and what a year it's been! While Fabric has taken away some of Power BI's limelight, we have certainly seen some impressive innovation on both fronts. This year, the "hardening" of Power BI files has all but been completed with PBIR soon becoming the default format, the TMDL view GA, the downloading of XMLA-edited files, web and desktop parity and so on. Naturally, GenAI has been at the forefront of most feature announcements across the industry this year, and Power BI and Fabric are no different with the advancements in Copilot, agentic features and MCP tools. All with the goal of improving your development, but your mileage may vary! If you'd like, you can read the official Power BI 2025 holiday recap: A decade of innovation and impact. Bring on 2026 and whatever that may bring. I dare say AI will continue playing a significant part!

So, onto the final set of highlighted articles in what's a pretty small edition. Marco Russo has written a couple of articles describing a couple of services/tools for your benefit: Introducing DAX Lib, the “app store” for DAX User-Defined Functions and Introducing DAX Optimizer Basic. And Chris Webb has again continued his series of helpful hints and tips on tuning Power BI Copilot performance with his article Power BI Copilot And Report Filters And Slicers.

Happy holidays everyone, and see you in the new year!

#337 - 9th December 2025

2025-12-09T00:00:00.00Z

It's a small edition this week, with only a couple of announcements on the official blog. There's one that's not really related to Power BI, but probably relevant to your wider data platforms: Assessing Your Azure Data Factory for Migration to Fabric Data Factory, a handy tool to help you assess whether you're ready to migrate your ADF workloads to Fabric based on the activities and functionality you use. The other announcement is that Drillthrough in Excel now supported for Direct Lake and DirectQuery Models, removing a limitation I'm sure many of you will now benefit from! Elsewhere this week, Chris Webb has shared information given by Celia Bayliss, describing how to Stop Power BI Copilot From Answering Questions From Report Visuals in certain scenarios where that is desired, Reid Havens has shared Two Simple Power BI Table Hacks: Invisible Tooltips and Zero-Width Labels, and Zoe Douglas has written about Deployment pipelines and Power BI semantic models with Direct Lake on OneLake tables and shared some information on how to get round some of the current limitations.

#336 - 2nd December 2025

2025-12-02T00:00:00.00Z

This week, the only announcement (which I'm sure a few of us expected was coming) is the Deprecation of Power BI Q&A. Since this is a long-standing piece of functionality and likely features in many reports, they're giving us a year (December 2026) to remove/migrate visuals before removing support. Better start migrating soon, because next December will come around quickly if this year's anything to go by! A semi-official blog this week comes from Jeffrey Wang, who's written about the powerbi-modeling-mcp tool introduced at Microsoft Ignite a few weeks ago in the blog Talk to Your Data Model: Introducing the Power BI Modeling MCP. I know I've highlighted a few of these blogs in the last few weeks, but even if you haven't read any of those, you should definitely read this one!

Elsewhere this week, Andrzej Leszkiewicz has shown an interesting way to achieve Responsive SVG Charts in Power BI Core Visuals, Gary Carpenter has presented a few videos about the Filter Pane or Slicer Pane in Power BI?, and Dániel Gábor Patkós has taken us through all we need to know about how to Refresh a Power BI PRO Semantic Model with your Fabric pipeline.

#335 - 25th November 2025

2025-11-25T00:00:00.00Z

A dainty edition this week, but there are some great bits of content. Starting with the announcements: it was shared in last week's feature summary, but the New Card visual in reports is now Generally Available with plenty of improvements over the "old" visual. Zoe Douglas has written an impact FAQ on the new card visual's transition from preview to GA which is worth scanning. Another highlight from last week's summary was the Image Visual Enhancements for Report Creators, for which we've now been given this slightly deeper-dive and walkthrough of some of the new functionality. Finally, we've been told about the GA of the TMDL Visual Studio Code Extension, and we now have a preview of Standalone Copilot in Power BI mobile apps: Chat with your data anytime anywhere.

On the community front this week, Marco Russo has cross-compared Model-dependent and model-independent user-defined functions in DAX and provided some useful suggestions regarding naming conventions of UDFs and parameters, Ed Hansberry has described how to Bind A Power BI Thin Report To A Local Model, and, if you're a bit lost with all the MCP Server talk over the last few weeks, then Eugene Meidinger has written a nice article about The Power BI modeling MCP server in plain English.

#334 - 18th November 2025

2025-11-18T00:00:00.00Z

Over the last week we've had quite a few official blogs, including the November 2025 Feature Summary, which now seems to be written (or at least posted) by Adam instead of Patrick. So that's new. Anyway, this month my favourite/the important updates are the Enhancements to Image visuals (including additional styling, state customization, and other formatting options), the Power BI Modeling MCP Server (Preview) which is contextualized in Mohammad's higher-level article Bringing Context Aware Intelligence to Power BI (which allows you to make changes to semantic models using natural language), and the deprecation of R and Python visuals in Embed for your customers solution (to be unsupported as of May 2026).

Other announcements this week include the fact that PBIR will become the default Power BI Report Format – Get Ready for the Transition. Interestingly, this doesn't mean that the `.pbix` file extension is going - rather that the internal structure of a PBIX will be the new PBIR format for all files going forward. On the Fabric side, there was an article about the Decoupling Default Semantic Models for Existing Items in Microsoft Fabric which started a few weeks ago. The models remain, it's just now easier to work on them independently from the underlying lakehouse/warehouse and delete them if desired!

Elsewhere this week, Gilbert Quevauvilliers has tested out Justin Martin's Power BI MCP Tuner Server (mentioned in the editorial of Issue 331) wondering Does it reduce capacity requirements?, Phil Seamark has written an interesting tidbit around The Impact of Primary Keys on DAX Query Performance in Power BI, and Reid Havens has spoken about the new Calendar features and functions in his video Speed Up Your Power BI Time Intelligence - with this Simple [DAX] Change.

#333 - 11th November 2025

2025-11-11T00:00:00.00Z

No official updates to provide this week. We have, though, been told about Advance your career in Data & AI with Microsoft Fabric Data Days starting on November 4th 2025 - 50+ days of live learning sessions hosted by experts where you can gain practical experience on Power BI and Microsoft Fabric. Also on the official blog, Christian Wade has written an article describing Power BI semantic models as accelerators for AI-enabled consumption detailing how Power BI models serve as key enablers of AI-powered insights.

Elsewhere this week, Chris Webb has written about Calling DAX UDFs From Power BI Copilot, Reza Rad showed us how to Get list of all Power BI DAX measures and their expressions and Meagan Longoria has described how to Check Power BI Bookmarks with Semantic Link Labs.

#332 - 4th November 2025

2025-11-04T00:00:00.00Z

Another short edition this week, with just a couple of official announcements - Preparing for the General Availability of the new Card visual, highlighting the main improvements to the new Card visual including rendering improvements, enhanced image support and expanded styling controls. We've also been told about the On-premises data gateway October 2025 release, which brings the compatibility in line with the latest version of Power BI Desktop.

Elsewhere this week, Gilbert Quevauvilliers has continued the series on Power BI load testing with the article Automating Load Testing Setting Up Your Fabric Lakehouse and Notebooks – Part 2. Meagan Longoria has described how to Check Power BI report interactions with Semantic Link Labs, and Chris Webb has written about Monitoring The DAX Queries Generated When The Power BI Copilot Index Is Built.

#331 - 28th October 2025

2025-10-28T00:00:00.00Z

A quiet edition but here are the highlights from the past week: Simplifying Reusable Logic in DAX with User-Defined Functions (UDFs) - DAX User-Defined Functions enable true modularity and reusability in Power BI semantic models, allowing developers to define logic once and reuse it across measures, calculated columns, and calculation groups without code duplication. MCP Server: DAX Performance Tuner - An MCP server that connects LLMs to Power BI models to systematically optimize DAX queries through performance analysis, research integration, and semantic validation to ensure optimizations return identical results.

TMDL View in Power BI: The Game-Changer for Semantic Model Management TMDL View provides a developer-friendly code editor for scripting and bulk-editing Power BI semantic models with IntelliSense, enabling reusable components, version control, and access to advanced features like perspectives.

And finally, Understanding The "Copilot Analyzed Only Part Of The Model Due To Its Size" Warning In Power BI Copilot - Power BI Copilot indexes text columns for natural language queries but has limits of 1,000 text columns and 5 million distinct values; the warning appears when these limits are exceeded and can be resolved by hiding key columns or adjusting the data schema.

#330 - 21st October 2025

2025-10-21T00:00:00.00Z

This week we've received the Power BI October 2025 Feature Summary, which announces the introduction of the new Power BI Controller in Power BI in PowerPoint (Preview), giving you a centralised way to manage and update all of your Power BI add-ins within a PowerPoint presentation. In your Power BI reports, it is also now possible to Automatically expand table columns to fill available space - eliminating the need for manual resizing and saving so much time! There are plenty more announcements in the feature summary, such as the latest updates in the Transition from Bing Maps to Azure Maps, the GA of Copilot to write DAX queries in DAX query view, and the rollout of the new Performance analyzer in web to see the load times of visuals.

Now onto the other official announcements. There is the introduction of A new Fabric REST API for connection binding of semantic models which gives you a way to programmatically configure and update your connection bindings, and we've also been told about the Improvements to the Report Copilot in Power BI which is available in the Power BI service.

Elsewhere this week, Jon Stjernegaard Vöge has shown how we can use RegEx in Power BI TMDL View for find and replace operations and provides some useful RegEx patterns, and Kenneth A. Omorodion has written a helpful overview on how to create Power BI Custom Calendars for Gregorian, Fiscal, Retail, 13-Month, and Lunar.

Barry Smart has published a 4-part series on The Data Product Canvas: a framework to help you build successful data products. You can read the articles here: Stop Building Data Products That Fail, Deep Dive into the Building Blocks, The Data Product Canvas in Action, and The Theory Behind The Canvas.

Finally, I would like to congratulate Ed Freeman (and of course, his wife Kirsty) on the birth of their first child, a baby boy called Elliot! Ed will be off on paternity leave for the next month; Jessica Hill & I will be guest-editing the Power BI Weekly Newsletter in his absence.

- Howard van Rooijen

#329 - 14th October 2025

2025-10-14T00:00:00.00Z

This week, we've had a few official articles. First of all, the "Prep Your Data for AI" feature is Now in the Power BI Service, meaning you don't need to switch tools if you're already working in Fabric. Microsoft is also now Empowering its customers to shape the future of data connectivity in Power BI and Microsoft Fabric by committing to drive their Power Query / Data Factory connector backlog based on ideas and votes by the community. Finally, the Power BI Report Server September 2025 Feature Summary, with a handful of visualisation & reporting updates, alongside highlighting that all on-premises reporting servises are being consolidated under Power BI Report Server as of SQL Server 2025. Over on the Fabric side, but still related to Power BI: we now have Service Principal Support in Semantic Link, so now notebooks can be automated without having to use your own credentials. And, presumably, you can automate some Admin workloads without having to be a permanent admin yourself.

Elsewhere this week, Philipp Lenz has shared a snippet to create a Power BI Date table in TMDL, Chris Webb has written an interesting article taking a First Look At Fabric Graph: Analysing Power BI Import Mode Refresh Job Graphs, and Carmel is back with her write-up of FabCon Vienna 2025: Day 2 and Day 3. Oh, and finally, my colleague Barry attended Big Data LDN 2025 - the hugely popular (free-to-attend) Data and AI conference that takes place in London every September. This year was dominated by AI agents, with speakers quick to highlight what amazing things their org was doing with AI. Barry has now attended 5 years on the trot and analysed the trend over that time and has some interesting insights - be sure to give his blog a read.

#328 - 7th October 2025

2025-10-07T00:00:00.00Z

This week, Zoe Douglas is back on the official blog providing a Deep dive into composite semantic models with Direct Lake and import tables, providing an in-depth walkthrough of this new, powerful public preview capability. Also on the official blog this week is the announcement around the Deprecation of Metric Sets in Power BI, which marks the end of a relatively short-lived feature (at least relative to Power BI's lengthy timeline).

Elsewhere this week, and also from the Microsoft team, Emily Lisa has joined Mike Carlo on the Power BI Tips YouTube channel to provide some Quick Tips on Web Modeling, which effectively (at least partially) enables one of the most requested features over Power BI's history - the ability to build Power BI reports on a Mac! Emily performs her whole demo on a Mac, showing the semantic model and report editing tabs working in conjunction for a functional dev experience in the web. Moving on - if you use Trello for your task management, Gil Raviv has shared an article and Open-Source code repository around Analyzing Trello boards in Power BI to get you up and running quickly. Finally, my colleague James Broome has written a couple of blogs around automating pipeline refreshes in a reliable and robust manner from Data Factory, either in Microsoft Fabric Data Factory Pipelines or in Azure Data Factory/Synapse Pipelines.

#327 - 30th September 2025

2025-09-30T00:00:00.00Z

As ever, a slightly quieter week after the excitement of FabCon. However, you'll find a few reflective blogs about it this week, and naturally I think you should take a look at my colleague Carmel's FabCon Vienna 2025: Day 1 breakdown, which is soon to be followed by Day 2 & 3. Great if you'd like a non-MSFT breakdown of the announcements and highlights from the conference. I also realised I missed the official DAX UDFs: Code once, reuse everywhere (Preview) blog last week, though there were plenty of other bits of content about UDFs that you probably saw. Onto the highlighted articles this week: Zoe Douglas has written about a feature of semantic models I didn't actually know about: KPIs. There's no way to create these directly through the Power BI Desktop UI, but you can use the TMDL view as Zoe explains. Very cool. Elsewhere, Phil Seamark has shared a write-up of a presentation of his, describing how to Create Custom Visuals in Power BI with GitHub Copilot. If you want a bit of a fun read about weird and wacky things you can do with Power BI custom visuals, then look no further!

#326 - 23rd September 2025

2025-09-23T00:00:00.00Z

After all the announcements at FabCon last week, we've had a bunch of "deep dive" blogs arrive from the Power BI team. I'll start with the miscellaneous ones - which are: TMDL view is now Generally Available, Deep Dive into Editing Semantic Models in the Power BI Service (also Generally Available) and Power BI in Teams – Content Shared in Teams Chats Now Opens a Dedicated Separate Window Within Teams. But now for the specifics. As discussed last week, DAX UDFs are now available. Read into it in the official blog DAX UDFs: Code once, reuse everywhere (Preview) or, alternatively, the SQLBI blog: Introducing user-defined functions in DAX. As you can imagine, various people have started playing around with them and creating useful functions already. For example, Davide Bacci has used them to create Easy UDF Chips, and Dane Belarmino has shown how to create a DAX UDF + HTML + SVG Custom Bar Chart. And you can find plenty more functions in the handy new DAXLIB project created by the SQLBI team. The other main announcement I mentioned last week was the new Calendar-based time intelligence: time intelligence, tailored (Preview) (Microsoft blog), which has also been elaborated by the SQLBI team: Introducing calendar-based time intelligence in DAX. Lots to sink your teeth into if you do a lot of data modeling!

#325 - 16th September 2025

2025-09-16T00:00:00.00Z

FabCon Europe has started this week, and the announcements are flowing in: make sure you keep track of the Fabric blog and Power BI blog for all the latest. On the Power BI side, we've just received the Power BI September 2025 Feature Summary, the most interesting updates (IMHO) being: the new Enhanced DAX Time Intelligence (Preview), giving you a method to create custom calendars over tables & columns in your data model and benefit from streamlined time-intelligence DAX functions; the preview of DAX User Defined Functions (both of which were teased at FabCon Vegas earlier in the year); and possibly the most exciting announcement is that we can now Download PBIX of XMLA altered semantic models. That is, no more "This file can't be downloaded yet" warnings in the Power BI Service (unless you're using incremental refresh, I'm sorry to say). There are plenty more announcements in that blog, such as the GA of Editing semantic models in the Power BI Service and Live editing Direct Lake semantic models with Power BI Desktop, but I'll let you browse the rest! The only separate announcement this week (so far) is the Deprecation of Power BI Integration within SharePoint Lists and Libraries.

Aside from FabCon and announcements this week, Mim is back talking about Power BI, taking a First Look at Incremental Framing in Power BI, explaining how it works and providing some benchmark use-cases, and Dane Belarmino has shown how we can use the Matrix visual in order to Visualize an Entire Year with a Power BI Calendar Heatmap.

#324 - 9th September 2025

2025-09-09T00:00:00.00Z

Nothing to announce this week. I should think the Power BI/Fabric teams are gearing up for FabCon Europe which is taking place next week. I won't be there, but a couple of my colleagues will be, so keep your eyes peeled on LinkedIn for some updates! On the community side this week, Gil Raviv has written about Sentiment Analysis in Power Query, discussing the deprecation of the old Power Query AI Insights feature, and offering a new solution using the Azure AI Language service instead. Davide Bacci is back sharing more design ideas, this time showing an effective way to Create a Launch Page using the HTML Content (Lite) visual. Finally, Koen Verbeeck has walked through using Workspace Identity Authentication for your Power BI Semantic Models in Microsoft Fabric, in his example using an Azure SQL DB as an example semantic model data source.

#323 - 2nd September 2025

2025-09-02T00:00:00.00Z

A pretty quiet week all round - no announcements and only a handful of content. Here are the bits of content I'd like to highlight this week: firstly, Marcus Wegener has shared a PoC he put together to show the art of the possible around creating Portable Power BI Dashboard Templates with Fabric Notebooks PoC, whereby you create a "template" report with placeholder objects that are then remapped to target objects using sempy, which eventually spits out a report. A great efficiency boost for cookie-cutter reports that you can easily use configuration to rebind to a new model. Chris Webb has written a new version of a blog written a couple of years ago around Performance Testing Power BI Direct Lake Models: Ensuring Worst-Case Performance after some recent improvements to the Direct Lake engine have changed how this works. Finally, Reza Rad has provided us with a useful Microsoft Fabric Org App VS Power BI App comparison video, describing the key differences and the implications of the new "org app" approach from an architectural perspective.

#322 - 26th August 2025

2025-08-26T00:00:00.00Z

Welcome to the 322nd edition of Power BI Weekly! A quiet week this week, with just the one article from the Power BI team. They're going into a bit more detail of one of the August feature summary announcements, which is the Semantic Model Refresh Templates in Power BI (Preview), which provide various ways to easily orchestrate your Power BI refreshed, including incremental refreshes, chaining dependent refreshe, and event-driven refreshes. All very common scenarios! Elsewhere this week, Matt Mair-Durrant has written an interesting article describing how to Narrate Your Data: Implementing Dynamic Alt-Text for Inclusive Reporting, taking advantage of the ability to populate Alt Text using a DAX formula much like you can for conditional formatting. Finally, Jon Stjernegaard Vöge has continued his series of experiments on Translytical Task Flows, providing a novel solution for Using Dropdowns for User Input in Translytical Task Flows.

#321 - 19th August 2025

2025-08-19T00:00:00.00Z

No announcements this week, just an introduction to the translytical task flow contest for you to showcase your examples of insight-to-action in your Power BI reports. No million-dollar prize money, sadly. But maybe some bragging rights amongst your colleagues! Elsewhere this week, on the technical side, Avi Singh has shared a useful tutorial on Visual Calc vs DAX in Power BI: Which One Should You Use? demonstrating the differences between the two approaches. Then, on the less technical side, but two complementary articles (IMHO), Kurt Buhler has written the article Five Minutes to Wow, talking about his journey into Power BI (via Tableau), the complexities around going from building quick and dirty reports to building enterprise reports, and where GenAI does and doesn't help (currently). Then I'd like to highlight an article written by Yugo Shimizu, translated (by Google) as Why aren't more users using Power BI correctly?. Yugo talks about the lack of understanding of what BI actually represents, the constant comparison to Excel and the problems that causes, the difficulty around actually "learning" Power BI enough to benefit from it at its fullest, and the over-reliance by some on AI. Well worth a read - just click the "translate" icon on your browser if you don't understand Japanese!

#320 - 12th August 2025

2025-08-12T00:00:00.00Z

A bit of a summer lull this week, even with the delivery of the Power BI August 2025 Feature Summary which doesn't contain a huge amount of updates. The most exciting (IMO) being the Support for org apps in Pro workspaces (Preview), allowing you multiple apps per workspace targeting different end-users, and comprising different combinations of workspace artifacts. As usual, we've also been given some Copilot updates, like Copilot now being available in Embedded Reports for SharePoint Online, and there have been a handful of modeling updates, such as support for Mirrored Azure Databricks catalog for Direct Lake in Power BI Desktop. Elsewhere this week, Marco and Alberto have given us a couple of interesting blogs on the SQLBI channel, discussing Sideways recursion in DAX calculation groups (which I'm sure keeps a lot of you up at night) and sharing some handy SUMMARIZECOLUMNS best practices.

#319 - 5th August 2025

2025-08-05T00:00:00.00Z

Let's start with this week's updates: first up is that the Standalone Copilot in Power BI will be turned on by default in September. As usual, you do get to manually override this if you don't wish to enable the "chat with your data" experience for whatever reason. We've also been told about the First-Ever Fabric Notebooks Competition for Power BI, while sharing a link to a new Notebook Gallery (which, it's no surprise to see, has been primed with a few notebooks from Michael Kovalsky (of Semantic Link fame)).

Elsewhere this week, there's some AI content on the SQLBI channel, with Kurt Buhler writing a great piece on AI in Power BI: Time to pay attention, describing the potential of MCP servers to complement your Power BI development flows. Also on the AI front, prathyoo has shared a useful prompt to allow you to Document PBIP report using AI. And on the DataViz front this week, specifically matrix visuals, Reid Havens has shown a neat way to achieve Dynamic Matrix Columns in Power BI: Show/Hide Based on Row Hierarchy, and Davide Bacci is back sharing a very impressive (and entirely native) Matrix Gradient visual, using some clever DAX logic.

#318 - 29th July 2025

2025-07-29T00:00:00.00Z

This week, Power BI turned 10! The tool has come a long way since then, cementing itself as one of the most consistently performing and popular BI products on the market. Kim Manis has shared a blog rounding up the key milestones: Marking 10 Years of Power BI: Celebrate with the Global Community, and there was a Power BI 10th birthday party! live-stream that Adam, Patrick and Amanda hosted on the Power BI YouTube channel, where they reflected, reminisced, heard from community members, heard from Power BI Leadership and had a general chat with those tuned in. Do take a look.

Back to reality, though, and the only announcement this week is that Cognitive services and Azure ML will be fully retired by September 15th, 2025 (in both Desktop and the Service), with the instruction being to move over to Synapse Data Science in Fabric - no surprises there! Elsewhere this week Ben Gribaudo has ventured into DAX blogging with an surprising insight about the Interplay Between No Selection Expression & Measures That Filter Calculation Groups, Bibiano Geraldo Mangue has written about the easy pitfall you can run into in the blog Great-Looking Dashboard, Bad Decisions: How Poor Data Preparation Misleads Your Analysis, and Ammar Asif has given us interesting an end-to-end walkthrough of a Power BI report about UK Laptop Imports vs Exports (2019–2025): Is It a One-Way Trade?

Finally this week, I wanted to highlight a guest blog post from Ray, a year 10 (9th grade) student who joined endjin for a week for some work experience. If you're interested in seeing what a "week in the life" of an endjineer entails, take a read of Ray's blog here: Hello World! I'm Ray and I'm doing work experience.

#317 - 22nd July 2025

2025-07-22T00:00:00.00Z

Plenty of great content this week. First of all, we've just been given the Power BI July 2025 Feature Summary which has a number of exciting updates. Unsurprisingly, there are loads of Copilot/AI updates generally extending the configurability and experiences of what you can do with Copilot in Power BI. Outside of AI, there are a number of great Reporting updates, including the ability to Influence sort for visual calculations (Preview), the General Availability of Field parameters, and the introduction of Organizational themes (Preview) (which apparently was announced last month on LinkedIn, which I missed). I'm particularly excited about the last update - an easy way to improve the consistency of reports across an organisation is a massive win.

There are a couple of relevant Fabric announcements this week. The first is the Introduction of access limits in a Fabric workspace - that is, you will no longer be allowed to have more than 1000 users+groups directly assigned to a workspace in Power BI/Fabric, which I can't imagine will be an issue for too many of us. The other announcement is the Sunsetting of Default Semantic Models – Microsoft Fabric, which has been a hot request from many in the community ever since Fabric was released. Default semantic models are those which are automatically created when you provision a Lakehouse in Fabric. There's currently no way to switch these off so they're not created in the first place, nor delete them when they're created. So this is welcome news!

Elsewhere this week, a couple of articles on more advanced Semantic Modeling features: Laura Graham-Brown has written about how Calculation Groups aren’t scary and Zoe Douglas has taken us through Perspectives in Power BI semantic models.

#316 - 15th July 2025

2025-07-15T00:00:00.00Z

Another short edition this week, with no announcements. Straight onto the highlighted articles, then. First of all, Bernat Agulló Roselló is back with an article describing how to Copy visuals between reports and prepare a report to make it bilingual, demoing how Tabular Editor, C# and the PBIP/PBIR file formats can work together to achieve something not possible in native Power BI Desktop. Elsewhere, Reid Havens has spoken about Everybody's Favorite Power BI & Fabric Topic: Licensing with guest Stephanie Bruno, providing a thorough summary of the relationship between user & capacity licences, as well as ways of calculating costs. Finally, Gilbert Quevauvilliers has started a new series on Power BI Load Testing with the article Why Should I Complete Power BI Performance Load Testing & Initial Setup. Outside of Power BI and Fabric, but very topical if you've been working with Lakehouses in Fabric or Direct Lake mode in Power BI, my colleague Carmel has published a video talking about How does Delta Lake work?

#315 - 8th July 2025

2025-07-08T00:00:00.00Z

A very small edition this week, but Copilot news has come to save the day. There have been several Updates to Copilot in Power BI: More ways to see, learn from, and ask about your report data, with the ability to ask data questions to your reports in "chat with your data" experience and the ability to ask ranking questions over your data. On the same topic this week, Chris Webb and Dan English have written about how you can piggy-back off AI instructions to achieve certain behaviours in Copilot: Chris has shown how you can use them to Prevent The Use Of Implicit Measures (which you may or may not agree with!), and Dan has shown how to use AI instructions to provide non-English language responses.

Also this week, I'd like to highlight a new blog series my colleague Barry has shared, all about DuckLake. You've probably heard about DuckDB - a high-performance analytical SQL engine over files and databases. However, it was not a Lakehouse format (or catalog) in and of itself, unlike what we get in a Synapse/Fabric Lakehouse which use a combination of Delta tables and a Hive metastore. Instead, DuckLake is a new integrated data lake and catalog format working over existing SQL Database Engines that is designed to provide enhanced performance and flexibility over existing Lakehouse+Catalog architectures. Very exciting - do check out Barry's blogs for more!

#314 - 1st July 2025

2025-07-01T00:00:00.00Z

A couple of official updates this week – Microsoft have announced a series of special events, learning opportunities and contests to Celebrate as Power BI Turns 10. We've also been told there are now Expanded entry points, access options, and functionality in Explore, which means the Explore feature is now available by default to users with view permissions, allowing more business users to perform ad hoc analysis on their data. The update also makes Explore more powerful: you can now launch it from both report visuals and Copilot generated visuals, and create visual calculations within your Explore analysis.

Now onto this week's highlighted content, Markus Ehrenmueller-Jensen has taken us through how to Become a Guardian of the Star Schema, helping us create create clean, high performance data models. Valerie Junk has shown how to use numeric range parameters for What if analysis in Power BI, Jon Stjernegaard Vöge has written an article Exploring User Input Options for Translytical Task Flows (Write-back in Fabric & Power BI). And finally, Carmel Eve has published a short talk: What is a Data Lakehouse?

#313 - 24th June 2025

2025-06-24T00:00:00.00Z

No technical updates this week, but Microsoft has found out that it has (yet again) been named a Leader in the 2025 Gartner® Magic Quadrant™ for Analytics and BI Platforms, and again being the data point furthest to the top-right of the chart. Very impressive stuff, and one to show your stakeholders if you ever need to justify the use of Power BI. Onto the highlighted articles in what is a very short edition: Bibiano Geraldo Mangue has taken us click-by-click through How to Display Power BI Reports in a Continuous Slideshow with Live Data, Justin Martin has shown how to use Workspace Monitoring alongside a KQL query to Identify Semantic Model Capacity Spikes, giving details of the user and DAX queries responsible for bad performance, and, lastly, Meagan Longoria has described how to Get Power BI Report Viewing History using Semantic Link Labs. Remember that Semantic Link (+ Labs) is not just to do with Semantic Modelling - it's scope is much, much broader!

#312 - 17th June 2025

2025-06-17T00:00:00.00Z

We've been provided with the Power BI June 2025 Feature Summary, which includes some greater flexibility when working with colleagues on pbix files with Sensitivity Labels applied, Power Query editing in the web for import models (Preview) (much like the experience for Dataflows that many of you will now be accustomed to), and that Field/Numeric Range parameters are now supported against Direct Lake tables (in Preview). Elsewhere on the announcements, Rui Romano has announced that we can now Open and edit any semantic model with Power BI tools - namely, using the TMDL view, you can now edit all properties of a semantic model without leaving Power BI Desktop - which has been a long time coming! Finally, Darren Gosbell has Released DAX Studio v3.3.0 including various usability enhancements and a new "High Contrast" theme.

#311 - 10th June 2025

2025-06-10T00:00:00.00Z

The main news from last week is that the Microsoft Fabric Community Conference Comes to Atlanta! from March 16-20, 2026. So no more Vegas (for now), which I'm sure some of you will be pleased with. Anyway, you have 9 months to book your tickets for what promises to be another conference full of new features and platform improvements! But back to the here and now - we've been told about the two new Copilot experiences now available, including the "Chat with your Data" experience (mentioned a few weeks ago) which has been fully rolled out, as well as the new feature whereby you can now get Copilot in Embedded Reports for Portals and Websites - so Copilot can follow your report into your custom application. Finally, we've been told about the SQL Server 2025 preview, which allows us to discover and experiment with What’s New in SQL Server 2025 Analysis Services.

Elsewhere this week, Soheil Bakhshi is back with Part 2 of Sharing Power BI Reports with External Users: Hands-On Guide to Setup and Sharing, stepping through Licensing, Admin settings, report permissions an discoverability. Andrzej Leszkiewicz has written two blogs about default slicer selections, outlining a means to achieve a Date Slicer with Default Selection, as well as covering Row Level Security → Default Slicer Selection, and this week we have Jon Stjernegaard Vöge taking us through the customary Translytical Task Flows use case, writing a Guide: Native Power BI Write-Back in Fabric with Translytical Task Flows (How to build a Comment/Annotation solution for Power BI).

#310 - 3rd June 2025

2025-06-03T00:00:00.00Z

No announcements this week. Though we've had a couple of blogs about the standalone Copilot experience in the Power BI Service, with Dan English giving us a quick run-through of the Standalone Power BI Copilot experience and Chris Webb demonstrating Finding Reports And Semantic Models Easily With Power BI Copilot Search. Elsewhere, Bas Dohmen has given us a nice overview and shown how we can Trigger Any Action from Power BI with Translytical Task Flows, Soheil Bakhshi has started a new series about Sharing Power BI Reports with External Users, with the initial blog: Part 1: Understanding the Problem and Core Concepts, and Augusto Rosa has written about New Passwordless Configuration: Configuring Single Sign‑On (SSO) from Power BI to Snowflake - a worthwhile read for any of you connecting your reports to Snowflake as a source.

#309 - 27th May 2025

2025-05-27T00:00:00.00Z

A couple of updates this week, which are the Power BI Report Server May 2025 Feature Summary (including a small set of updates from the March Power BI Desktop release) and the On-premises data gateway May 2025 release. On the community side this week, we have a number of interesting articles. Firstly, Ben Gribaudo has made us (or at least me) aware of Power Query's default rounding behaviour, which differs to the high-school rounding most of us will have learned. Philipp Lenz has shown how to create and retrieve Descriptions of key performance indicators/measures and how to visualize/display them in reports. Teo Lachev has written about the good, bad and the ugly in the article First Look at Fabric Translytical Task Flows. Finally, Dan English has written a blog highlighting a small feature from the latest feature summary that may have gone unnoticed, and that's the ability to Paste List of Values into Slicer. So you can now copy some values from Excel and paste them straight into the new List and Button slicers!

#308 - 20th May 2025

2025-05-20T00:00:00.00Z

Welcome to the 308th edition of Power BI Weekly! Sorry it's a bit late, but it's a bigger one than usual, what with Microsoft Build this week. If you're interested in all the announcements across Microsoft's services, check out the Book of news which has a big old article describing all the main announcements and developments. Naturally, there's been plenty of announcements in the Power BI and Fabric offerings too. Those announcements relating to Power BI are as follows:

Translytical task flows (Preview) - the feature that was teased last year at Ignite. The native ability to "write-back" (and much more) in Power BI.
The next era of Copilot in Power BI: Chat with your Data (Preview) - new dedicated Copilot experience in the Power BI/Fabric interface (yes, you needed another one). See also: Extracting deeper insights with Fabric Data Agents in Copilot in Power BI
Updates to Fabric Copilot Capacity - now available to F2 capacities and above, and a new admin setting being switched "on" by default.

And on the topic of Copilot, Jeffrey Wang has written a fascinating article about LLMs and DAX: Where Things Stand Today after last week's article highlighting the improvement of Copilot's DAX capabilities. The whole article's worth a read, but Jeffrey's ever-wise words stand-out at the end: "...remember that the smartest move you can make is the same one Power BI experts have always preached: invest in a clean, well-structured semantic model."

Now, moving on. We've also received the Power BI May 2025 Feature Summary this week! The main thing they've drawn our attention to is the new Power BI section of the Fabric Roadmap (preview), which gives us a clear view as to what features are being worked on behind the scenes (across all of Fabric, not just Power BI). There are also a bunch more Copilot & AI updates, including a new "Prep data for AI" experience in Power BI Desktop that you may have seen appear over the last few days. Other announcements this month include Persisted sorting for field parameters (Preview), and a Deep dive into selection expressions for calculation groups, given you further control about how Calc Groups behave when multiple items are selected, or no item is selected.

Finally this week, I wanted to highlight a number of new endjin blogs. First of all, a couple of blog posts from my brilliant colleague Paul, who wrote a blog taking us through How to Build Mobile Navigation in Power BI, and also provided a A Creative Walkthrough of our Titanic Power BI Report, a brilliant demonstration how some professional design thinking can level up your reports. For those bridging into Data Engineering (especially on Fabric), my colleague Carmel has written a trio of blog posts touching on some of the modern foundations, with the articles: What is a Data Lakehouse?, How do Data Lakehouses Work? An Intro to Delta Lake and What is the Medallion Architecture? Finally, my (ex) colleague Liam has now moved on to pastures new, but not without writing the article Retrospecting on my career at endjin. If you're interested in learning more about what a role at endjin looks like, please take a read of Liam's blog. We wish him all the best!

#307 - 13th May 2025

2025-05-13T00:00:00.00Z

A couple of announcements this week: there have been some improvements to Fabric Copilot to help write DAX queries, which "include overall correctness of the DAX queries, deeper understanding of how to utilize DAX row and filter context, and how to use semantic model features such as inactive relationships" and also "increased stability, meaning similar user requests phrased in different ways create similar or identical DAX queries", which is brilliant news. DAX has enough complexity as it is, without you having to worry about understanding whether Copilot has spat out garbage or not. Hopefully we're inching towards DAX outputs that can largely be trusted - but always test your measures! On the Fabric side this week, we've been told about the general availability of Shortcut cache and on-prem gateway support, which is great news for those of you with hybrid/cross cloud data connections. And great news for your bill-payers too. Aside from announcements this week, lbendlin (sadly I can't figure out their name) has described a Streamlined process for appending multiple files with similar structure in a folder in Power Query, and Chris Webb has demonstrated how to Find Events Linked To A Specific Power BI Visual In Fabric Workspace Monitoring.

#306 - 6th May 2025

2025-05-06T00:00:00.00Z

A handful of announcements this week. The first is the new Fabric SQL Database Integration with Power BI Desktop, a new preview feature to connect to Fabric SQL Database via the OneLake Catalog. We've also been told about the (inevitable) deprecation of Datamarts in favour of Fabric Data Warehouse, and we've been told that the XMLA endpoint is going to switch to be read/write by default next month to reduce confusion, particularly when using external tools to integrate with your semantic models.

Elsewhere this week, Jon Stjernegaard Vöge has written a detailed article describing Migrating from Power BI Premium Capacity to Fabric Capacity (Cross Region) – What you should know (a situation I'm sure many of you find yourselves in), and Chris Webb has written about Documenting Power BI Semantic Models With Fabric Data Agents, showing a mostly impressive set of responses from the Agent.

#305 - 29th April 2025

2025-04-29T00:00:00.00Z

Just the one official update that arrived just after last week's edition was compiled, and that's the IntelliSense improvements to DAX query view, which highlights the new autocomplete/pre-selection behaviour to make it writing your code more intuitive (and therefore less frustrating!) Now onto this week's highlighted content: on the data viz side, Davide Bacci has shared a brilliant new Deneb-based Top N & Others Pie Chart visual (Pie Charts aren't always ineffective!), Bas Dohmen has show some smarter Quadrant Charts in Power BI with the latest update of Power BI Desktop, and, outside of Data Viz, Koen Verbeeck has written Part 2 of Power BI Mistakes to Avoid series, sharing some useful information from DAX structure through to security considerations.

#304 - 22nd April 2025

2025-04-22T00:00:00.00Z

Quite a few official articles this week. First of all, the FabCon Las Vegas keynote is now available to watch on YouTube, which includes most of the announcements that were shared at FabCon (if you haven't had a chance to read all the materials). Something that was highlighted in the keynote was Microsoft's strong performance in market analyses, and this week we've learned that Microsoft named a Leader in The Forrester Wave: Business Intelligence Platforms, Q2 2025 - congrats (as ever) to the Power BI team.

Probably more interesting news to this readership is the Power BI April 2025 Feature Summary. It's (unsurprisingly) quite a light one this month, but here are some highlights: there's now Ad hoc calculation support for data questions now available in read mode for Copilot, the Mobile layout auto-create functionality is now GA, and there are some model-layout-related updates to the Live edit of semantic models in Direct Lake mode with Power BI Desktop Preview feature. And on that last point, Zoe Douglas has provided a Deep dive into Direct Lake on OneLake and creating Direct Lake semantic models in Power BI Desktop for those that want to explore deeper.

Finally, I'd like to highlight that I'll be presenting the talk Microsoft Fabric: Too Many Options - Which One(s) Should I Choose? at the Oxford Microsoft Data Platform Group on 7th May (both online and in-person). There's "choice-overload" in Fabric - so I'll be implementing the same underlying data pipeline using different Microsoft Fabric item types and comparing and contrasting them for you. Would love to see you there (either physically or virtually!)

#303 - 15th April 2025

2025-04-15T00:00:00.00Z

This week, the Power BI team has taken us through the New mapping and location analytics capabilities in Microsoft Power BI, highlighting that Azure Maps visuals can now be "published to web", alongside new built-in capabilities to add data-bound reference layers and Path layers. Elsewhere, the Cosmos DB team has announced the Power BI Integration with vCore-based Azure Cosmos DB for MongoDB Now in Public Preview, making it easier than ever to create real-time dashboards over your operational data in this source.

From the community side this week, Antriksh Sharma has described how to Implement Fill Up and Fill Down in DAX (somewhat more complicated than Power Query, but this is dynamic!), and Gilbert Quevauvilliers has shown how Using TMDL in Power BI Desktop to bulk replace saved hours of time. Finally, again from the Power BI blog, there was a fun Sanofi Power BI Data Visualization Competition: An Inspiring Initiative Focused on the Paris Olympic Games featuring a number of appealing, interactive dashboards over the open Olympic data. Some great design inspiration found in those examples - go take a look!