Power BI essentials — Data Viz Module 8 | LeadAfrik Public Economics Hub

Power BI is Microsoft's answer to Tableau. For most enterprises, the deciding factor is licensing, not features — if your organisation already has Microsoft 365 E3 or E5, Power BI is essentially free. That distribution advantage means Power BI is now the most-deployed BI tool in finance functions globally, even though Tableau remains the brand favourite for analyst-built dashboards.

The three layers

Power Query (M language): connect to sources, clean and shape data. ETL done in a Excel-like recipe interface that records every step.
Data model: define relationships between tables. Star schema (one fact table joined to multiple dimension tables) is the standard pattern.
DAX (Data Analysis Expressions): the formula language for measures and calculated columns. Looks like Excel formulas, behaves more like SQL.

The star schema

Power BI rewards a clean star schema: one fact table (transactions, sales, events) connected by foreign keys to multiple dimension tables (date, product, customer, region). Avoid flat 'one big spreadsheet' designs. Performance, DAX correctness, and refresh times all hinge on the data model being properly normalised.

DAX, in three formulas

text

// 1. Simple measure
Total Revenue = SUM(Sales[Amount])

// 2. Time intelligence — same period last year
Revenue LY = CALCULATE([Total Revenue], SAMEPERIODLASTYEAR(Date[Date]))

// 3. Filtered context — revenue from VIP customers only
VIP Revenue = CALCULATE([Total Revenue], Customer[Tier] = "VIP")

// 4. % of total
Revenue % of Total = DIVIDE([Total Revenue], CALCULATE([Total Revenue], ALL(Sales)))

Master CALCULATE, SUMX, FILTER, ALL, and SAMEPERIODLASTYEAR and you can write 80% of finance-team DAX.

Power BI vs Tableau

Tableau wins: faster exploratory analysis, better visual polish out of the box, the chart-as-art aesthetic.
Power BI wins: integration with the Microsoft stack, DAX for complex financial measures, lower cost in Microsoft shops.
Both: real-time dashboards, scheduled refresh, embedded analytics, mobile apps.

When to learn which

If your target employer is a Microsoft shop — most banks, insurers, government — Power BI is the practical choice. If you're building public-facing dashboards or you want hire-ability across more places, Tableau still has the slight edge. The good news: knowing one makes learning the other a weekend.

The reporting layer

Power BI Service (the cloud component) is where dashboards live and where row-level security gets enforced — important for finance dashboards where Region East shouldn't see Region West's numbers. The data model is built in Power BI Desktop on your laptop; published to the service; refreshed on a schedule. This separation of authoring and serving is mature and works.

Exercise

Download Power BI Desktop (free). Connect to any public CSV — a World Bank indicator file works well. Use Power Query to clean it, build a simple star schema (even if there's only one dimension table, e.g. date), write three DAX measures (total, YoY growth, % of total), and build a one-page report with one chart per measure and a slicer for date.