Data and AI Training

Home | Prices | Contact Us | Courses: Power BI - Excel - Python - SQL - Generative AI - Visualising Data - Analysing Data

Power BI Overview

This is a whirlwind tour to introduce you to Power BI and help you decide if you want to invest time and effort in learning how to use it well.

What is Power BI?

Power BI is a (proprietary) business intelligence tool from Microsoft. It is the most popular BI software on the planet (according to Microsoft). Here is Microsoft’s Power BI marketing page.

Desktop and Service

Power BI has two parts:

power bi desktop screenshot
Power BI Desktop as of March 2025

power bi service screenshot
Power BI Service as of March 2025

Power BI Desktop Components

Power BI Desktop stitches together several components:

Power BI Desktop has several views and editors. You will get to know these well. These are:

Import, shape and clean data in the Query Editor

The Query Editor is a thing of wonder! It allows us to import our data for all sorts of different sources (database, CSV, Excel, …), clean and shape our data to make it ready for visualisation and analysis. It has a user interface so we can apply the transformations that we need. Under the covers, it writes a language (called M) so these steps can be automated and data refreshed on demand or on schedule

power bi query editor
Power BI Desktop’s Query Editor.

Generative AI models, such as ChatGPT, can help a lot with Power Query and M.

power bi query editor
ChatGPT conversation to solve Power Query problem.

Design the visuals and report pages in the report view

A Power BI report has one or more pages. Each page contains one or more visuals. A visual is an interactive chart - the report consumer will be able to filter, drill into detail on the chart.

Charts on a page can interact with each other. For example, selecting a particular category on a bar chart can filter the data shown by other visuals on the page to that category only.

The report view has the report canvas and several panes.

power bi report view canvas
Power BI Desktop’s report view canvas.

The panes are:

Other panes are also available and occasionally useful.

power bi report view panes
Power BI Desktop’s report view with panes on the right. From right to left these are Data, Build, Format and Filters.

The visuals gallery contains about 30 different types of visual. We can group these into:

power bi visuals gallery
Power BI Desktop’s visual gallery

View data in the Table View

In the table view, you can see all the data in the model (and filter, sort the data).

power bi table view
Power BI Desktop’s Table View

Build a semantic model in the Model View

A semantic model is a copy of all the data imported by the Query Editor. It usually has several tables. But its much more tha a simple dataset - the semantic model also includes the metadata / meaning of the model added by the designer:

Data (semantic) modelling is the fundamental challenge! The analytical power of Power BI comes from arranging several tables with relationships. Relationships match columns into two tables that have the same domain of values. This allows report builders to choose any set of columns from any table in the data model and place on to a visual and that visual will work properly - no need for any Excel XLOOKUP or SQL JOIN statements.

power bi model view
Power BI Desktop’s Model View.

Write and test measures in the DAX Query View

The DAX query view is a game changer for learning, understanding, writing and testing DAX. It is an alternative to writing DAX in the formula bar. It is a proper editor (using the codebase of the very popular VSCode). It encourages structured thinking about DAX problems rather than hack-it-and-hope approach.

power bi dax query view
The DAX query view helps us write, test and understand DAX.

The “Quick queries” functionality writes some starter example DAX.

power bi dax quick queries
The “Quick Queries” context menu in Power BI Desktop’s DAX Query View.

Edit the report and model TMDL and the TMDL view

TMDL stands for Tabular Model Definition Language. It is a domain-specific language to define and manage Power BI models in Power BI. It is text-based and human readable (in YAML syntax) so can be edited and version-controlled.

You can edit the TMDL either in

If you save a Power BI file as Project (PBIP) format, this creates also two subfolders:

power bi tmdl view
Power BI Desktop’s TMDL View.

vscode tmdl
Edit TMDL in VSCode.

DAX is the calculation and modelling language

The two things that you need to know about DAX before you start your journey.

Visual Calculations and Generative AI tools, both discussed later, can make DAX less difficult.

Visual Calculations

Visual Calculations are promoted as “DAX made easy”. They provide a point and click method of writing many typical calculations in DAX that would be more difficult to write as a measure on a visual. There are some pros and cons compared to measures.

power bi visual calculations example
Visual Calculations Example.

Power BI Enterprise Features

Enterprise features are more important to large development teams building complex sematic models. Microsoft have been retrofitting enterprise features into Power BI Desktop and Service over the last few years. These include:

Use AI models to help with Power BI challenges

In order of importance, AI Models can help with:

  1. DAX
  2. M (the language of PowerQuery)
  3. data modelling advice (import an image of the semantic model into the AI model)
  4. TMDL (but we are all learning how to use this best)
  5. report layout and design (Copilot only)

Two options

External AI tools work well once we provide some context about the model to the AI tool. This could be an image of the model view or an equivalent written description of the model (in terms table and column names, star schema arrangement, facts and dimension tables etc).
Here is an example exercise and a shared ChatGPT conversation.

Skills needed for mastery of Power BI

Several skills are needed to master all of Power BI.

Typically, one person does not have all of these skills. A team is often split into sub-teams:

Other tools are available as well as Power BI Desktop

There are other tools as well as Power BI Desktop that allow report builders to see further under the covers. These are usual free or freemium licenses. They include:

There are available as standalone and also available as extensions to Power BI Desktop.

power bi visuals gallery
This Bravo model analysis shows column sizes and highlights columns that are not used on any visual on any report so could possibly be removed

The shortest history of the evolution of Power BI

Here is the shortest form of the long history of Power BI. It start with the heritage that goes back further than you may expect.

The Power BI story proper starts more the a quarter century ago.

power bi desktop 2016
Power BI Desktop from the October 2016 blog post.

Why learn Power BI?

Recommendation: Use Power BI as part of your data analyst’s armoury

My recommendation - use Power BI as one tool of many to help solve data challenges. Other tools may include:

Resources to help learn Power BI

There are a lot of great resources available, for example:

A short list of learning resources here.

power bi query editor
A Getting Started in Power BI playlist on YouTube. Direct Link here.

How to get started

  1. Go to a Power BI meetup, either in-person or virtual.
  2. Grab some interesting data, either from work or public data.
  3. Sell your Mac, buy a PC, then download Power BI Desktop (it’s free).
  4. Have a go. Us the learning resources above. And remember “5 minutes to wow”.

Power BI is part of Microsoft Fabric

Power BI is now one component Of Microsoft Fabric. This is Microsoft’s end-to-end data everything (BI , data science, data engineering) offering

Fabric is purely Software as a Service (SaaS). Access through a web browser

The heart is a data lake named OneLake. Think of this as a collection of large files. (Data is in Delta-Parquet format). Many tools can read / write from the OneLake

fabric landing page
Microsoft Fabric landing Page.

Here is Microsoft’s elevator paragraph. Microsoft Fabric is an enterprise-ready, end-to-end analytics platform. It unifies data movement, data processing, ingestion, transformation, real-time event routing, and report building. It supports these capabilities with integrated services like Data Engineering, Data Factory, Data Science, Real-Time Intelligence, Data Warehouse, and Databases.

fabric stack
Microsoft Fabric Architecture Stack.

fabric workspace example
Fabric Workspace Example.

fabric task flows
Fabric Task Flows.

fabric sql endpoint
Fabric SQL Endpoint.

fabric Python Notebook
Fabric Python Notebook.