Power BI Advanced Series · Module 1 — Lesson 1 · by Raushan Ranjan, MCT

Why the Advanced Track Is Different

You already know what Power BI is. You may have built a few reports. The beginner series teaches you to drive the car. The advanced track teaches you how the engine works — so that when your 10M-row import times out, your DAX measure returns wrong numbers, or your published report shows stale data, you know exactly why and exactly what to fix.

Lesson 1 builds the mental model you will use for every subsequent lesson: what data analysis actually is, the four types every analyst performs, the role you are training for, and the six-step workflow Power BI is built around.

Quick Answer Data analysis is the process of converting raw data into decisions. Power BI structures that process into six stages: Acquire → Transform → Model → Analyse → Visualise → Share. Every tool in Power BI maps to one of these stages. Understanding which stage you're in at any moment is the difference between a productive session and an hour of confusion.
  • Four analysis types: Descriptive, Diagnostic, Predictive, Prescriptive — in order of increasing insight value
  • Power Query handles stages 1–2 (Acquire and Transform)
  • The data model and DAX handle stages 3–4 (Model and Analyse)
  • Power BI Desktop and Service handle stages 5–6 (Visualise and Share)
Data Analysis with Power BI
From raw data to business decisions — the Power BI pipeline.

The Four Types of Analysis — What the PL-300 Expects You to Know

These four types are not theoretical. The PL-300 exam tests them directly, and interviewers use them to filter candidates. More importantly, every stakeholder request you receive in your career maps to one of these four types — and knowing which one tells you which Power BI tool to reach for.

TypeQuestion It AnswersPower BI Feature UsedEnterprise Example
Descriptive "What happened?" Bar charts, line charts, matrices, card visuals Total sales by region last quarter — summarising what occurred
Diagnostic "Why did it happen?" Drill-through, Key Influencers visual, decomposition tree Revenue dropped 12% in Q3 — drilling into product category, region, and sales rep to find root cause
Predictive "What will happen?" Forecasting on line charts, Python/R visuals, Azure ML integration Projecting next quarter's demand based on 3-year seasonal trend
Prescriptive "What should we do?" What-if parameters, scenario analysis, goal seek Modelling the impact of a 10% price increase on margin across product lines
PL-300 tip: The exam often presents a business scenario and asks which analysis type it represents. Memorise the question each type answers — that is the fastest classification shortcut.

The Data Analyst's Role — Five Responsibilities

The PL-300 is structured around the Data Analyst's responsibilities. These map directly to exam task areas:

Prepare Data
Connect, clean, and shape data in Power Query. This is the PL-300's largest topic area — ~25% of exam questions.
Model Data
Build relationships, create hierarchies, define calculated tables. Star schema fluency is required.
Visualise Data
Choose the right visual, configure interactions, format for clarity and accessibility.
Analyse Data
Write DAX measures, use time intelligence, create what-if scenarios.
Deploy & Maintain
Configure workspaces, set up RLS, manage refresh schedules, monitor gateway health.

The Six-Stage Workflow — Your Mental Map for Every Task

Every action you take in Power BI fits into one of these six stages. When you feel lost, ask yourself: "Which stage am I in right now?" That question alone will point you to the right tool.

StageToolOutputCommon Mistake
1. AcquirePower Query — Get DataRaw query connected to sourceUsing ODBC instead of native connector — loses query folding
2. TransformPower Query EditorClean, typed, shaped tableDoing transformations in Excel before import — breaks repeatability
3. ModelModel view / RelationshipsStar schema with defined cardinalitySkipping this stage — DAX breaks on missing or incorrect relationships
4. AnalyseDAX measures in Report viewCalculated KPIs and metricsUsing calculated columns where measures are needed — inflates model size
5. VisualiseReport canvasInteractive report pagesUsing too many visuals per page — cognitive overload, slow render
6. SharePower BI ServicePublished dataset + workspacePublishing without configuring refresh or RLS — stale data and security gaps
⚠️ 3 Mistakes at the Starting Line
  1. Starting in Report view before transforming data. The canvas is the most inviting part of Power BI Desktop. But dragging fields before your data is clean and your model is defined means you will rebuild everything after you fix data issues. Always complete Power Query → Model before touching the canvas.
  2. Connecting to data with the wrong mode and never revisiting it. The connectivity mode (Import / DirectQuery) is set at connection time. Changing it later requires rebuilding queries. Decide before you connect: how large is this dataset? How fresh does it need to be? That decision takes 30 seconds up front and saves hours later.
  3. Publishing to "My Workspace" and sharing from there. My Workspace is a personal sandbox — it cannot be assigned a Premium capacity, cannot use deployment pipelines, and its datasets cannot be certified. For any report shared with others, create a dedicated app workspace from the start.
✅ Key Takeaways — Lesson 1
  • The four analysis types (Descriptive, Diagnostic, Predictive, Prescriptive) map every stakeholder request to a specific Power BI tool
  • The six-stage workflow (Acquire → Transform → Model → Analyse → Visualise → Share) is your mental map for every Power BI session
  • Power Query handles stages 1–2; the data model and DAX handle 3–4; Desktop and Service handle 5–6
  • Never start on the report canvas before your data is clean and your model is defined
  • The PL-300 exam is organised around the Data Analyst's five responsibilities — every exam topic maps back to one of them
🧠 Check Your Understanding

Q1. A financial analyst asks: "Based on last year's seasonal patterns, how many units should we order for Q4?" Which analysis type is this, and which Power BI feature would you use to answer it?

Show Answer

Predictive analysis — it is asking "what will happen?" based on historical patterns.
Power BI feature: the Forecast option on a line chart (Analytics pane → Forecast) for simple projections, or an Azure Machine Learning integration / Python visual for more sophisticated models. What-if parameters can also model order quantity scenarios (which edges into Prescriptive).

Q2. You are about to connect Power BI to a 500MB SQL Server table that updates every 30 minutes and must reflect data no older than 1 hour. Which connectivity mode do you choose, and why?

Show Answer

DirectQuery. The 500MB size is below the practical Import limit, but the 1-hour freshness requirement rules out Import (which on shared capacity refreshes at most every 30 minutes, and refresh scheduling has overhead). DirectQuery queries the source directly on every interaction, guaranteeing data is always current. Trade-off: report performance depends on SQL Server query speed.

Q3. In which stage of the six-step workflow does DAX belong, and what is the key mistake that makes DAX measures return incorrect results even when the formula is written correctly?

Show Answer

DAX belongs to Stage 4 — Analyse.
The key mistake is skipping Stage 3 (Model). DAX measures use filter context propagated through relationships. If the relationship between your fact table and dimension table is missing, has wrong cardinality, or uses the wrong cross-filter direction, DAX formulas will return incorrect totals, ignore filters, or throw blank results — even when the formula syntax is perfectly correct.

Next in Module 1 Lesson 2: Connecting to multiple data sources simultaneously — SQL Server, SharePoint Excel, Azure SQL, and Dynamics 365 in one model. Read Lesson 2 →