Most of my talks are demo heavy, or take a feature further than the documentation does. Usually there is a measurement or a finding in there that makes you think again about whatever I am talking about.
Fabric Apps SpeedrunAbstractFabric Apps let you build a working application inside Fabric from one or two prompts. They are a cannon: you can blow a hole into the world with it, or take your own leg off. In this lightning talk I will show how Fabric Apps work underneath, what they can cost, where they sit next to Power BI, Power Apps and Power Automate, and how to think about them so they bring progress instead of an unmaintainable time bomb.
Fabric pod kontrolou: itemy, náklady a kapacitaAbstractAdministering a Power BI estate was never simple, even back when we only had a handful of item types. Microsoft Fabric takes it a level further, and keeping the runaway machine under control is no joke. In this session we will show you how to find out what is really happening in your tenant. Who creates new items and where, who uses them, what each operation costs, and which workload is throttling your capacity right now. We start with Microsoft's built-in tools, go on through community toolkits, and finish with solutions built on Power BI, Fabric REST APIs and semantic models. We will go from simple no-code options to writing your own, so Fabric stops being an expensive black box.
Když Power BI začne zapisovatAbstractPower BI reads, analyses and visualises data rather than creating it. With Translytical Task Flows and User Data Functions that role is starting to shift. A report no longer has to be something you only look at. It can write data back, start a workflow, or interact with other items in Microsoft Fabric. Scenarios like these were possible before through the Power Platform, but part of that functionality is moving into Fabric itself. In this session I want to walk you through the typical use cases, show how to set the whole workflow up, and explain what to watch for in architecture and in CU consumption. We will also look at how large a difference the wrong write mechanism makes, and how to reach sources outside Fabric in the edge cases.
Direct Lake pod kapotouAbstractDirect Lake is a fine addition to the Power BI storage modes we already know, but it wants a different approach than Import or DirectQuery. In this (not only) benchmark session I will show what came out of my own testing of Direct Lake in Microsoft Fabric. We will look at how the design of the underlying Delta tables affects performance, what happens when new data arrives, how to keep frequently used data quickly available, what running it actually costs, and how Direct Lake works in general. The goal is to introduce Direct Lake properly, so nothing surprises you once it is in production.
PragueCZ
WUG standalone session
User Data Functions End-to-End: Ve Fabricu i mimo nějAbstractFabric User Data Functions (UDF) and Translytical Task Flows let you solve advanced interactivity and write-back directly in Power BI, without needing Power Apps. In this session we take a broader look at UDFs: how they work, what they cost in CUs, and where they help or get in the way. We will build them, and we will get around the restrictions that aim them at Fabric items only. In a live demo I will show concrete use cases for talking to external sources such as Azure SQL Database or an Excel file on SharePoint, and how to wire all of it neatly into a Power BI report.
Dataflow Gen2: Evolution Power Query and the price for comfortDataflow Gen2: Evoluce Power Query a cena za pohodlíAbstractWith Fabric, Microsoft introduced a new generation of Dataflows. It brings the familiar comfort of visual transformations, and underneath it offers entirely new storage options and a tuned compute engine. Higher performance arrives with a new price tag. In this session we look at whether it makes sense to deploy Dataflow Gen2 for real, and what its capacity (CU) consumption is next to the other tools in Fabric. The aim is to answer the question that matters: when migrating to Gen2 pays off, when staying on the proven Dataflow Gen1 is the wiser move, and when leaving the transformations in Power BI Desktop is still the best answer.
OnlineEN
Super User Meeting
Ingesting and Transforming Data in Fabric: Optimizing with Capacity Units in MindAbstractFabric gives you several ways to ingest and transform data, each with its own Capacity Unit cost. Depending on the tool you pick, the spend can vary a lot. In today’s session I will focus on Fabric Python notebooks, not PySpark, which are often the most economical choice for small to medium data engineering tasks. However, they call for a mostly pro-code approach. That’s why I will show how to pull data from external sources like APIs, land it in a Lakehouse, and perform transformations with Polars while keeping Capacity Units low.
BrnoCZ
WUG standalone session
Power BI a Fabric v praxi pro jednočlenný datový týmAbstractThis one is aimed at small data teams, or rather at the situation where the whole team is you. As a one man show you do most of the work in Power BI, heavy transformations and writing M in Power Query included, and you may want to cut refresh times or bring the overall cost and CU consumption down. I will show you how to move towards a data engineer's role in Fabric, with data preparation for Power BI in mind, what you need to know, and how to build your first notebook and data pipeline step by step.
Power Query for DevelopersAbstractIn this session, we will look together at what Power Query is and how to approach it as developers. We will write M language, cover basic syntax, evaluation and types, and learn how to write custom steps to transform data. We will go through built-in functions and data connectors, and we will also look at how to build our own data connectors, for example, for REST API sources.
How does Power Query evaluate your queriesJak Power Query (M) vyhodnocuje dotazy a proč je výkon někdy nekonzistentníAbstractThe session explains how Power Query M really processes a query, why the steps are not evaluated the moment they are defined, and how lazy evaluation works in practice. You will learn when data gets read more than once, how Buffer can help with that, and why the same query behaves differently depending on context, data source, or the way it is used.
PragueCZ
WUG standalone session
REST APIs jako datový zdroj v Power BIAbstractIn this session we look at using a REST API as a data source in Power BI. We start with the basics of Web.Contents() and go through the most common ways of authenticating against an API. You will also learn how to handle pagination efficiently when loading large volumes of data. The session is built on practical examples, tips and tricks for optimising queries included. The emphasis is on Power Query and the M language.
BrnoCZ
WUG standalone session
REST APIs jako datový zdroj v Power BIAbstractIn this session we look at using a REST API as a data source in Power BI. We start with the basics of Web.Contents() and go through the most common ways of authenticating against an API. You will also learn how to handle pagination efficiently when loading large volumes of data. The session is built on practical examples, tips and tricks for optimising queries included. The emphasis is on Power Query and the M language.
DAX Query: Next-Level Data Source InsightsAbstractLost in a sea of reports with no clear view of your data sources? Sure, you could dig through them manually or use the Power BI REST API, but neither tells the full story. DAX Query changes that, and it’s super easy! With its latest capabilities, you can reveal every table, M transformation, SQL statement, and file path powering your reports. In this session, we'll explore how to implement these insights effortlessly using many Microsoft Power components, giving you complete visibility and control over your data sources.
Modeling in Power BI: How does it work in real lifeModelování v Power BI: Jak to chodí v reálném světěAbstractMost of us have heard the basics of modelling in Power BI, and what an ideal data model should look like, a beautiful star schema for instance. Reality tends to be different. You work with assorted sources that were never designed for efficient modelling. In this session I will share what that looks like in practice: how to stitch various Excel files and other semi-structured sources together into a data model that works and is fast.
PragueCZ
WUG standalone session
Power Query: Když GUI nestačíAbstractWe explore what to do where the Power Query interface runs out. Writing your own functions, working with loops, and more advanced integration such as APIs, authentication and basic pagination. We will show how to get past the limits of the graphical interface with M code and your own connectors, which open up new ways of working with a source.
Master Power Query - Introduction to M languageOvládni Power Query - Úvod do jazyka MAbstractThe aim is to pick up the basics of the M language. We go through the elementary principles, and how we actually “write” M by clicking the ready-made functions hidden behind the buttons in Power Query. Then we look at how writing my own M code makes a dataset easier to work with and faster. We will go through the use cases where knowing the basics of M makes working with a source far easier.
Master Power Query - Introduction to M languageOvládni Power Query - Úvod do jazyka MAbstractThe aim is to pick up the basics of the M language. We go through the elementary principles, and how we actually “write” M by clicking the ready-made functions hidden behind the buttons in Power Query. Then we look at how writing my own M code makes a dataset easier to work with and faster. We will go through the use cases where knowing the basics of M makes working with a source far easier.
Invite me
Want one of these at your event?
A conference, a user group, or a session inside your company. Czech or English, in person or remote. Tell me roughly what your audience does and I will say which one fits, or write you a new one.
One email when something new goes up. Nothing else, and one click to stop.
I use Google Analytics to see which posts get read and to tell returning readers from new ones. It sets cookies. Nothing from Google loads unless you agree. What is collected