5.3.1. Power BI Components and Choosing a Visual
Choosing the right chart is crucial for clarity:
| Visualization | Best For | Avoid When |
|---|---|---|
| Bar/Column Charts | Comparisons between categories | More than 10-12 categories |
| Line Charts | Trends over time | Non-continuous data |
| Pie Charts | Proportions of a whole | More than 5-6 slices |
| Scatter Plots | Correlation between two numeric values | Non-numeric data |
| Cards | Single KPI values | Multiple related values |
| Tables | Detailed lookup | Executive summaries |
💡 First Principle: Every visual encodes data as a visual property — length, position, angle, area, colour. Pick the visual whose property matches the question being asked, and the answer becomes obvious without the viewer doing arithmetic.
Beyond the Basics
| Visualization | Best For | Avoid When |
|---|---|---|
| Matrix | Cross-tabulated detail with expandable row and column groups, like a pivot table | You need a single headline number |
| Treemap | Hierarchical composition, each rectangle sized by its value | Values are close together — areas are hard to compare precisely |
| Card | One headline number, shown large | You need to show a trend or a comparison |
| KPI | One value measured against a target, with trend | There is no target to measure against |
| Gauge | Progress toward a goal within a known range | The range is unbounded |
| Slicer | Letting the user filter the whole report page by selecting values | You want a fixed, non-interactive view |
A slicer is not a chart. It is an on-canvas filter control: the reader picks a region or a date, and every other visual on the page responds. When a question describes users "narrowing down" or "filtering the report themselves", it is describing a slicer, not a visualization.
Reasoning Tool — read the verb in the question:
| The question says... | Reach for |
|---|---|
| "compare across categories" | Bar / column chart |
| "trend over time" | Line chart |
| "proportion of the whole", "percentage of total" | Pie or donut chart |
| "breakdown within a hierarchy, sized by value" | Treemap |
| "expand and collapse groups", "like a pivot table" | Matrix |
| "correlation between two numbers" | Scatter plot |
| "one single metric" | Card |
| "against a target" | KPI |
| "let users filter" | Slicer |
Desktop, Service and Mobile: Who Does What
Power BI is three products, and questions turn constantly on which one a task belongs to.
| Component | What it is | What you do there |
|---|---|---|
| Power BI Desktop | Free Windows authoring application | Connect to sources, shape data in Power Query, build the model, design report pages |
| Power BI Service | The cloud platform (app.powerbi.com) | Publish, create dashboards, share, set scheduled refresh, manage workspaces and licences |
| Power BI Mobile | Phone and tablet apps | View and interact with published content — authoring is not available |
Reasoning Tool: if the verb is build, the answer is Desktop. If it is share, schedule, pin or govern, the answer is the Service. Dashboards in particular exist only in the Service.
Power Query: Where the Data Gets Cleaned
Power Query is the data-preparation engine inside Desktop (and behind dataflows in the Service). It connects to the source and applies transformation steps — renaming columns, changing types, splitting fields, removing errors, merging tables — recorded in order so the whole sequence replays on every refresh.
The important idea is that this shaping happens before the data reaches the model. By the time a visual queries it, the cleaning is already done and repeatable.
Workspaces
A workspace is the collaboration boundary in the Service. My Workspace is personal and shares with nobody; a shared workspace is where a team co-owns datasets, reports and dashboards, with roles controlling who may edit and who may only view. Content is published from a workspace to a wider audience as an app.
