Describe an Analytics Workload on Azure · Topic 4 of 4
Power BI Visualization
DP-900 Power BI: Desktop, service and mobile apps, building a report, data models with relationships and measures, and choosing the right visual for the data.
- 6 min
- Desktop builds. The service shares. Mobile apps view. Report Builder makes paginated reports.
- A report is interactive and built on one model. A dashboard is a page of pinned tiles.
- Measures are calculated with DAX. Power Query prepares the data.
What it is
Microsoft Power BI is the tool that turns prepared data into interactive reports and dashboards. Analysts build them, and decision makers explore them to find trends and answer questions.
Power BI is also a core component of Microsoft Fabric, so it reads the same data as the other workloads.
How it works
The Power BI components
| Component | What it is | Used for |
|---|---|---|
| Power BI Desktop | A free Windows application | Building models and reports |
| Power BI service | The online software as a service | Publishing, sharing, dashboards and collaboration |
| Power BI mobile apps | Apps for phones and tablets | Viewing reports and dashboards on the move |
| Power BI Report Builder | A tool for paginated reports | Print ready, fixed layout reports |
A common flow is to build in Desktop, publish to the service, and share through a workspace.
Reports and dashboards
A report has one or more pages of visuals, all built on one data model, and it is interactive: selecting a bar filters the other visuals. A dashboard is a single page of tiles, pinned from one or more reports, that gives an at a glance view. Dashboards are made in the service.
Data models
A data model is made of tables and the relationships between them. Good models follow the star schema, with fact tables in the middle and dimension tables around them, which keeps reports fast and simple.
- Relationships link tables by a key column, usually one to many, and decide how filters flow.
- Measures are calculations, such as total sales, written in the formula language DAX and calculated as you interact with the report.
- Calculated columns add a new column to a table.
- Hierarchies group fields, such as year, quarter and month, so users can drill down.
Data is prepared with Power Query, which cleans and reshapes it before it enters the model. In current Power BI documentation, the model that reports use is called a semantic model, which was previously called a dataset.
Power BI can import a copy of the data into the model for speed, or use DirectQuery to send queries to the source each time so the results are always current.
Choosing a visual
The right visual depends on the question being answered.
| Question | Visual |
|---|---|
| Compare values across categories | Bar or column chart |
| Show change over time | Line chart |
| Show parts of a whole, with few categories | Pie or donut chart |
| See the relationship between two numbers | Scatter chart |
| Show a place based pattern | Map |
| Show one key number | Card |
| Compare to a target | KPI or gauge |
| Show hierarchical proportions | Treemap |
| Show exact values by two categories | Table or matrix |
| Let users filter the whole page | Slicer |
A worked scenario
An analyst connects Power BI Desktop to a warehouse. In Power Query they remove blank rows and fix date formats. In the model they relate a Sales fact table to Customer, Product and Date dimension tables and create a measure for total revenue. The report page has a card for total revenue, a line chart of revenue by month, a bar chart by region, and a slicer for the year. After publishing to the Power BI service, the analyst pins the key visuals to a dashboard for the sales director, who reads it on a phone.
Building a report in steps
- Open Power BI Desktop and choose Get data to connect to a source such as a database or a file.
- Use Power Query to remove errors, rename columns and set data types.
- In the model view, create relationships between the fact and dimension tables.
- Add measures with DAX, for example total revenue and revenue for the same period last year.
- Place visuals on the report page, then add slicers and a title.
- Publish to the Power BI service, share it, and schedule a data refresh if the data was imported.
Filtering and exploring
Users explore a report in several ways. A slicer is a visible control for filtering. Filters can apply to one visual, one page or the whole report. Selecting one bar cross filters the other visuals, and a hierarchy lets users drill down from year to quarter to month. These interactions are the reason a report is more useful than a static chart.
Sharing and security
In the service, content lives in a workspace, and people are given a role that controls what they can do. You can share a report with colleagues, or package content as an app for a wider audience. Row level security restricts which rows each viewer can see, for example so that a regional manager sees only their own region's sales, even though everyone opens the same report.
Import or DirectQuery?
| Import | DirectQuery | |
|---|---|---|
| Data location | Copied into the model | Left in the source |
| Speed | Fast | Depends on the source |
| Freshness | As of the last refresh | Current at each interaction |
| Typical use | Most reports | Very large or constantly changing data |
What to remember for the exam
- Desktop builds. The service shares. Mobile apps view. Report Builder makes paginated reports.
- A report is interactive and built on one model. A dashboard is a page of pinned tiles.
- Measures are calculated with DAX. Power Query prepares the data.
- Star schema models give fast, simple reports.
- Pick the visual from the question: line for trend, bar for comparison, pie for part of a whole, map for place.
A line chart is the usual answer for change over time, and a bar or column chart is the usual answer for comparing categories.
How the exam asks about this
- "Which tool is used to author reports on a Windows PC?" Power BI Desktop.
- "Show how sales change month by month." A line chart.
- "A single page of pinned tiles from several reports." A dashboard.
- "The language for calculating measures." DAX.
Common mistakes
Confusing a report with a dashboard. Reports have pages and filter interactively. Dashboards are single pages of tiles.
Using a pie chart for many categories. It becomes unreadable. A bar chart compares many values better.
Building a model without relationships. Visuals then cannot filter one another correctly.
Thinking Power BI Desktop is cloud based. It is a desktop application. The service is online.
Check yourself
Which Power BI component is used to build a report on your PC?
Power BI Desktop.
Which visual best shows a trend over time?
A line chart.
What is a dashboard in Power BI?
A single page of tiles pinned from reports, giving an at a glance view.
Topic 4 of 4 in Describe an Analytics Workload on Azure
Test yourself on Describe an Analytics Workload on Azure
The full Microsoft DP-900 question bank has 503 questions, with timed practice exams and an explanation for every answer.
This study guide is independent and is not affiliated with, endorsed by, or sponsored by Microsoft. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.