Customize this training

Get a certificate — $34.29$12

Data Visualisation and Dashboards with Power BI and Tableau

Data Visualisation and Dashboards with Power BI and Tableau

Choose the right chart, model your data, write basic DAX and Tableau calculations, then build, publish and check an accessible dashboard in both tools.

What you will learn

  • Define the audience, decision and KPIs of a dashboard before building anything.
  • Choose the right chart for each question and avoid misleading visuals.
  • Shape data into a clean star-schema model and check it before charting.
  • Write basic DAX measures and Tableau calculated fields, including LOD expressions.
  • Build and publish a Power BI report with a scheduled refresh and controlled access.
  • Build and publish a Tableau dashboard with actions, device layouts and permissions.
  • Apply storytelling and accessibility rules and run a pre-launch checklist to avoid common mistakes.

Modules

  1. Start with the decision, not the chart
  2. Choosing the right chart
  3. Data modelling basics
  4. DAX and Tableau calculations
  5. Build and publish in Power BI
  6. Build and publish in Tableau
  7. Storytelling, accessibility and common mistakes

Course facts

Estimated study time
about 60 minutes
Final assessment
41 questions · Pass mark 75%
Available in
English, Français, العربية
Price
Free. Optional verifiable certificate: $12
Last updated
Published by
AMAADOR ACADEMY

Who this course is for

Analysts, reporting and finance staff, managers, marketers and job seekers who need to turn data into dashboards that people actually use.

Prerequisites

Comfort with spreadsheets (tables, sums, filters). No coding needed. Free Power BI Desktop (Windows) and Tableau Public or a Tableau trial are useful for the exercises.

Course content

1. Start with the decision, not the chart

Most failed dashboards fail before the first chart: nobody knew who would use them or what they would decide. In Power BI and in Tableau the work starts on paper, not in the tool.

  1. Audience — Who opens it: an executive on a phone, an analyst at a desk, a shop manager every morning?
  2. Decision — What will they do differently after reading it: reorder stock, call a customer, cut a cost?
  3. Questions — Write 3 to 5 plain questions: "Are we on target this month?", "Which region is behind?"
  4. Frequency — Daily, weekly or monthly? This sets the data refresh.
  5. Action — What happens when a number turns red, and who owns it?

Then choose the type of product. A dashboard is a single screen for monitoring. A report has several pages for exploring. An analysis answers one question once. Many requests called "dashboard" are really a one-off analysis.

Four answers you need before you open Power BI or Tableau.

Try it: take a report you receive today. Write the decision it supports in one sentence. If you cannot, nobody needs the report in its current form.

2. Choosing the right chart

Chart choice follows the question, not taste. Ask what the reader must see: a comparison, a trend, a share of a whole, a spread, a relationship or a place.

QuestionBest choiceAvoid
Which is bigger?Bar chart (sorted)3D bars, radar
How does it change over time?Line chartPie per month
What share of the total?Stacked bar, or pie with at most 5 slicesPie with 12 slices
How is it spread?Histogram, box plotAverage alone
Are two measures related?Scatter plotTwo lines on two axes
Where?Map, only if location mattersMap when a bar would do
Exact values?Table, cardChart with no labels

In Power BI these are the clustered bar, line, donut, scatter and matrix visuals; in Tableau you drag fields and use Show Me, then override it when you disagree.

Case: the 1986 Challenger launch decision. Edward Tufte argued (in Visual Explanations, 1997) that the pre-launch charts did not show O-ring damage against temperature for all flights in one view, so the link with cold weather was hard to see. He proposed one scatter plot of all flights. Other scholars dispute parts of his account (for example which temperature to plot and what the engineers already knew), but the lesson holds: show the relationship the decision depends on in a single view.

3. Data modelling basics

Both tools work best with a star schema: one central fact table with events and numbers (sales lines, tickets, shipments) surrounded by dimension tables with descriptions (customer, product, date, region). Microsoft recommends it for Power BI in its star schema guidance on Microsoft Learn.

Table typeHoldsExample columns
FactOne row per event, numeric measures, keysOrderLineID, DateKey, ProductKey, Quantity, Amount
DimensionOne row per thing, descriptive attributesProductKey, Name, Category, Brand
  1. Decide the grain — what one fact row means ("one order line"). Write it down.
  2. List the dimensions — who, what, where, when.
  3. Relate them one-to-many — one product has many sales lines. The "one" side is the dimension.
  4. Add a date table — a full calendar with one row per day, marked as the date table in Power BI, so time calculations work.
  5. Hide technical keys from report users.

Filters flow from the dimension (the "one" side) to the fact table. Selecting "Category = Shoes" in a slicer filters the sales lines of shoes only.

4. DAX and Tableau calculations

DAX (Data Analysis Expressions) is the formula language of Power BI. Learn five ideas and you can build most business reports.

  1. Measure, not column — a measure is calculated at the moment of use, in the context of the chart. A calculated column is stored per row. For sums, ratios and KPIs use measures.
  2. Aggregate — Total Sales = SUM(Sales[Amount])
  3. Safe division — Margin % = DIVIDE([Profit], [Total Sales]) returns blank instead of an error when the denominator is zero.
  4. CALCULATE changes the filter — Sales 2025 = CALCULATE([Total Sales], 'Date'[Year] = 2025)
  5. Time intelligence — with a marked date table: Sales LY = CALCULATE([Total Sales], SAMEPERIODLASTYEAR('Date'[Date])), and TOTALYTD for year to date.

Understand filter context: each cell of a chart is calculated with the filters that apply to it (the row's product, the slicer's year). A measure gives a different answer in every cell without you writing a loop.

NeedPattern
Growth vs last yearDIVIDE([Total Sales] - [Sales LY], [Sales LY])
Count of customersDISTINCTCOUNT(Sales[CustomerKey])
Share of all productsDIVIDE([Total Sales], CALCULATE([Total Sales], ALL(Product)))

5. Build and publish in Power BI

Power BI Desktop is the free authoring tool for Windows. Microsoft ships a new version every month, so menus move: use the names below and check Microsoft Learn when something looks different.

  1. Get data — connect to Excel, CSV, SQL Server, SharePoint and many other sources.
  2. Transform data — open Power Query: set types, remove duplicates, merge or append tables. Each step is recorded and replayed at every refresh.
  3. Model view — check relationships (one-to-many, single direction), mark the date table, hide keys, write measures in a dedicated table.
  4. Report view — drag fields to visuals; use cards for KPIs, slicers for filters.
  5. Interact — add drill-through pages, tooltips pages and bookmarks for a guided story.
  6. Theme — apply a colour theme (a JSON file) so every report in the team looks the same.
  7. Check performance — run Performance Analyzer to find slow visuals.
Storage modeWhen to choose it
ImportDefault. Fast, data copied into the model, needs scheduled refresh
DirectQueryVery large or near-real-time data; speed depends on the source
Direct LakeMicrosoft Fabric lakehouse or warehouse data

6. Build and publish in Tableau

Tableau builds a dashboard from worksheets. Each worksheet is one view; the dashboard arranges several views on one canvas. Tableau releases new versions by year and quarter (2026.1, 2026.2 and so on), so check the release notes for your version.

  1. Connect — Excel, CSV, databases or a published data source. On the Data Source page add relationships between tables.
  2. Live or extract — a live connection queries the source each time; an extract (Hyper file) is a fast snapshot refreshed on a schedule.
  3. Build sheets — drag dimensions to Rows or Columns and measures to the marks. Use Color, Size and Label sparingly.
  4. Create the dashboard — choose a size (Automatic or a fixed size), add sheets, use tiled layout containers for alignment and floating objects only for small items such as a logo.
  5. Add actions — Filter (click a bar to filter other sheets), Highlight, Go to Sheet, URL.
  6. Device layouts — create Tablet and Phone layouts so mobile readers do not scroll sideways.
  7. Titles and tooltips — write titles that state the message; edit tooltips to show only what helps.

7. Storytelling, accessibility and common mistakes

People scan a dashboard in seconds. Design so the message is found in five seconds, then supported by detail.

  1. Top left: the answer — KPIs with target and trend, in the first place the eye goes (left to right in English and French, right to left in Arabic reports).
  2. Middle: the why — two or three charts that explain the KPIs: trend, breakdown, ranking.
  3. Bottom or drill-through: the detail — tables for people who need exact rows.
  4. Titles state the finding — "Returns rose 12% in March, driven by shoes", not "Returns by month".
  5. Annotate — a short note beside the spike explains it (a promotion, a stock-out).
  6. Use white space — align objects, leave margins, keep one font family.
LayerQuestion answeredExample visual
SummaryHow are we doing?KPI cards with target
ExplanationWhy?Trend line, ranked bars
DetailWhich exact items?Table, drill-through page

For Arabic audiences, set the report or dashboard so the reading direction and number format suit the reader, and test the final page with real Arabic labels, which are often longer or shorter than English.

The full lesson continues in the interactive course on this page.

Sources and references

Educational course. Menus, licences and features of Power BI and Tableau change often: check the current official documentation before you rely on a feature. Not affiliated with, endorsed by or sponsored by Microsoft or Salesforce. Power BI is a trademark of Microsoft; Tableau is a trademark of Salesforce. Dashboards that show personal data must follow the privacy law that applies to you (GDPR, Moroccan Law 09-08 and others).

All courses · AI & Data - Free courses, certificate from $7 | AMAADOR ACADEMY