⚡ A scroll-through explainer

How Your Power BI Dashboard
Actually Works

A plain-English guide for everyone who's ever looked at a dashboard and wondered:
"Where did this number come from?"

The complete Power BI tech stack — click any layer
☁️
Layer 06 · Share & govern
Power BI Service
Who sees what, when, and how → explore ↓
Distributed
📊
Layer 05 · What humans see
Visualisation
Charts, tables, slicers, drill-downs → explore ↓
Visible
🧮
Layer 04 · Business logic
DAX Measures
Live formulas that recalculate instantly → explore ↓
Computed
🕸️
Layer 03 · The brain
Semantic Model
Tables, relationships & structure → explore ↓
Modelled
🍳
Layer 02 · Clean & shape
Power Query
Transforms raw data into clean tables → explore ↓
Transformed
🗄️
Layer 01 · Raw ingredients
Data Sources
SQL, Excel, APIs, cloud apps → explore ↓
Source

↑ Data flows bottom → top  ·  Click any layer to explore

Scroll to explore
Section 01 · Data Sources

The Raw Ingredients —
Power BI Doesn't Create Data

Power BI connects to your data wherever it already lives. Before any chart, any formula, any dashboard — the data has to come from somewhere.

Think of Power BI as a chef, not a farmer. It doesn't grow the ingredients — it sources them from wherever they live, and turns them into something useful.
🗄️
SQL Server
Relational databases
📊
Excel / CSV
Files & spreadsheets
☁️
SharePoint
Microsoft 365
🔵
Azure
Cloud data lakes
⚙️
SAP / Oracle
Enterprise systems
🌐
Web APIs
REST / JSON feeds
📋
Dataverse
Power Platform
📈
Google Sheets
Third-party files

⚠️ The Golden Rule

Garbage in, garbage out. If the source data is messy, incomplete, or wrong — Power BI will show you a prettier version of that mess.

Data quality at the source matters more than any formula you write downstream.

Section 02 · Power Query

Before the Fancy Charts:
The Messy Kitchen

Every great dashboard starts with ugly, raw data. Power Query is where the transformation happens — invisibly, automatically, every time.

Step 01
🚚
Ingredients Arrive
🗄️ SQL Server
📊 Excel
🌐 Web API

Raw data arrives from all over — databases, spreadsheets, websites. It's messy. Duplicates, nulls, wrong formats.

Step 02
🔪
The Prep Work
Remove duplicates
Fill in missing values
Fix date formats
Rename confusing columns
Filter out test data

Power Query is your kitchen prep station. Every transformation is recorded — like a recipe. Change an ingredient? The whole recipe re-runs.

Step 03
🍽️
Ready for the Chef
DateProductSales
01/01/24Widget A$1,200
02/01/24Widget B$840
03/01/24Widget A$2,100

Clean data loads into the model. Now the DAX measures can do their magic.

💡 Why this matters for you: If a number looks wrong on your dashboard, the problem is usually HERE — in the prep kitchen, not the chart itself. Always check Power Query first.

Section 03 · Semantic Model

The Brain Behind
Your Data

Tables don't talk to each other by magic. Relationships wire them together — and when you click a filter, the whole model responds instantly.

📅
Data Model Group Chat
3 tables online

The Star Schema — Your Data Model's Blueprint

Every Power BI model has a hierarchy. Here's how it works.

Sales Table
The Team Lead
(Fact Table)
📅
Calendar
Specialist
📦
Product
Specialist
👤
Customer
Specialist
🌏
Region
Specialist

"The Team Lead doesn't store details themselves — they ask their specialists. That's a Star Schema."

📐 What's Cardinality?

One calendar date → many sales. One customer → many orders. This "one-to-many" relationship is called Cardinality.

Think of it like one teacher → many students. The teacher (date) doesn't repeat themselves, but the students (sales) can refer to that same teacher many times.

Section 04 · DAX Measures

The Live Recipes
Behind Every Number

A DAX Measure is not a number stored in a spreadsheet cell. It's a live recipe that recalculates the moment you touch a filter — like a barista who makes your coffee fresh every time, instead of pouring from a pot that's been sitting there since morning.
Total Sales
🧾
The Order
"Show me total sales"
👨‍🍳
What it does
Adds up every sales row that's currently visible — after all active filters are applied
The Magic
Change the date slicer? It re-adds. Change the region? It re-adds again. Instantly.
📈
Year-over-Year Growth
🧾
The Order
"How much did we grow vs last year?"
👨‍🍳
What it does
Makes two coffees — one for this year, one for last year — then calculates the difference
The Magic
Uses Time Intelligence — it knows what "last year" means even if your filter says "2024"
🍕
% of Total
🧾
The Order
"What % does each product contribute?"
👨‍🍳
What it does
Temporarily ignores the product filter to get the grand total, then divides your slice by it
The Magic
REMOVEFILTERS inside CALCULATE — strips away the current product filter so the engine sees the grand total, not just your filtered slice

🔍 What's Context Transition?

Imagine you're sitting at a restaurant table. When a calculation needs totals, the waiter doesn't look at all tables — instead, he takes your table number and turns it into a filter, so the kitchen knows to calculate only for your table.

That conversion — from "this row" to "a filter" — is Context Transition. It happens whenever CALCULATE is called inside a row context (like inside SUMX or a calculated column). DAX converts each row into an equivalent set of filters before evaluating the expression.

Section 05 · Visualisation

The Layer Humans
Actually See

Charts, tables, KPI cards, slicers, maps — this is the UI. It sits on top of all the logic below. And here's the thing most beginners miss:

Visuals don't change the data. They change the filter context. If a number looks wrong — the bug is almost never here.

🔍 Where to look when something's wrong

Number incorrect? → Check the DAX measure
Data missing? → Check Power Query
Wrong rows showing? → Check relationships in the Semantic Model
No data at all? → Check the data source connection

The visual is almost never the problem. It's a window — not the room.

Section 06 · Power BI Service

From Your Laptop
to the Whole Team

Power BI isn't one app — it's three. Knowing which does what will save you a lot of confusion.

🍳
Power BI Desktop
= The Home Kitchen
  • ✅ Free to use
  • 🔨 You build (cook) reports privately
  • 🚫 Nobody else can see your work yet
  • 💻 Lives on your local computer
💬 "Like writing a Word document on your own laptop. Only you can see it."
🍽️
Power BI Service
= The Restaurant
  • ☁️ Cloud-based — open for business
  • 👥 Others can view your dashboards
  • 💳 Requires Pro licence to share ($10/user)
  • 🔄 Enables scheduled data refresh
💬 "Like publishing to SharePoint so the whole team can read — and comment — in real time."
🚚
The Gateway
= The Delivery Truck
  • 🔗 Bridges cloud ↔ on-premise databases
  • 🕐 Runs 24/7 in the background
  • 🔄 Refreshes data on schedule
  • 🛡️ Keeps data secure inside your network
💬 "The truck picking up fresh ingredients every morning so the restaurant menu always shows today's specials."

🙋 Your First Questions, Answered

The most common cause is a filter you didn't notice — check the slicers at the top of the page. The second most common cause is the data hasn't refreshed yet. Look for a "Last updated" timestamp somewhere on the dashboard.
If you're a Viewer, you can interact with filters and drill into data — but you can't change the chart itself. Think of it like a restaurant menu: you can choose what to order, but you can't redesign the menu. Only the report builder (Editor) can do that.
Check for a "Last refreshed" or "Data updated" label on the dashboard. Most Power BI reports refresh once or twice a day via a scheduled refresh — like a newspaper delivered each morning. For real-time data, you'd need a special setup.
This is usually a DAX measure doing something clever — like removing duplicates, applying special logic, or calculating a weighted average instead of a simple sum. It's not a bug, it's intentional. Ask the report builder what the measure is designed to count.
Yes! Most visuals have a "..." menu in the top right corner. Click it and look for "Export data". You'll get a CSV or Excel file of the underlying data. Note: some reports have this disabled for security reasons.

⚡ What Happens When You Click a Filter?

Every click triggers a chain reaction across the entire model — here's what happens in under a second.

1
👆 You click "2024" on the date slicer
Your filter intent enters the model
2
📅 Calendar Table receives the filter
"Boss wants 2024 only. Marking all other years as invisible."
3
🔗 Filter flows through relationships
Calendar tells Sales, Product, and Customer tables to hide 2023 data
4
🧮 DAX measures recalculate
Total Sales, YoY Growth, % of Target — all re-cooked fresh with only 2024 data
5
📊 Charts update instantly
Every visual on the page re-renders. All of this happens in milliseconds.
The whole chain completes in under 1 second.
That's the power of a well-built data model.

📖 Words You'll Hear in Meetings

Plain-English explanations for the Power BI jargon everyone throws around.

Slicer
The filter buttons at the top of the dashboard
Like the genre filter on Netflix — click it and everything else updates to match your choice.
Measure
A live calculation, not a stored number
Like a barista making your coffee fresh — not pouring from a pot that's been sitting there all day.
Dimension
A "who/what/when/where" table
The Calendar, Product, and Customer tables are dimensions. They give context to the raw numbers.
Drill-through
Zooming in on one specific thing
Right-click a bar chart → "Drill through" → see all the detail rows behind that one bar. Like clicking into a folder.
Cardinality
How tables are related
One teacher → many students. One date → many sales. That "one-to-many" relationship is cardinality.
Refresh
The data newspaper delivery
Once or twice a day, the Gateway fetches fresh data and updates all the numbers. Like your morning paper arriving.
Star Schema
Hub and spoke data design
One central "fact" table surrounded by specialist tables. Like a team lead (Sales Table) surrounded by their team (Date, Product, Customer).
Row-Level Security
Personalised data views
Different people see different rows. A regional manager only sees their region's data. The dashboard is the same — the data filtered is different.
The Takeaway

Now You Know What's
Behind the Numbers

📊

Your charts are the tip of the iceberg. The real work — cleaning, modeling, calculating — happens in the layers below, invisible to most viewers.

🔗

Every click you make sends a message through the data model, triggering a chain reaction that updates every visual on the page — instantly.

🧮

The numbers aren't stored — they're calculated fresh, every time, just for you and your current filter context. That's what makes Power BI so powerful.