What you will be able to do
- Set up a visualization widget: dataset, chart type and axes
- Format widget titles with rich text and live @ references, and inspect the query a visualization runs
- Add text widgets and image widgets, and choose the right image source
1.Visualization widgets
An AI/BI dashboard is built by placing widgets on the canvas, and the widgets display data from the datasets the dashboard defines. A visualization widget turns one dataset into a chart. When you place one on the canvas, the Visualization Configuration panel opens on the right.
Because of that default, the first thing to do on a dashboard with several datasets is to check the Dataset drop-down. Then set up the chart:
1. Select a dataset from the Dataset drop-down. 2. Add filters (optional): click Show filters to apply a static filter or a parameter. 3. Choose a visualization type from the Visualization drop-down. 4. Define the x-axis and y-axis fields. Drag fields to reorder them, and use the display name to rename an axis label.
In draft mode you can open the widget's kebab menu and click View visualization query. This shows the SQL the widget actually runs, which is narrower than the dataset query: it returns only the data needed to draw that chart. Any filters you have applied are already in it. Parameter values are not filled into the SQL; they are listed at the top of the modal instead.
Checkpoint 1 of 6· Check yourself
In draft mode you open View visualization query on a bar chart. What does the modal show?
The query belongs to the widget and differs from the dataset query. Parameter values are listed at the top of the modal, not substituted into the SQL.
“This query is specific to the widget and only returns the data necessary to render that visualization.”Source: docs.databricks.com
Widget titles and descriptions support rich text: fonts, colours, alignment, links, tables and markdown. Type @ in a title or description to insert a live reference:
- A filter reference shows the viewer's current selection. For example, "Sales for @region" reads "Sales for West". - A field reference shows a live aggregated value, such as "Total revenue: $1.2M", and updates when filters change.
Checkpoint 2 of 6· Exam question
A dashboard must show order volume from the `orders` table alongside marketing spend from a separate `marketing_spend` table, displayed as two independent widgets on the same page. How should the analyst configure the underlying data for this?
Correct answer: A — Create two separate datasets, one query against `orders` and one against `marketing_spend`, and assign each widget to the matching dataset.
- A. Each dataset is defined from its own query or table, and a dashboard can hold multiple datasets, so pairing one dataset per source table with its own widget keeps the two sources independent.
- B. Joining the tables into one dataset would force both widgets to share a single combined result set, which is unnecessary when the two widgets need to chart unrelated metrics separately.
- C. A filter option only supplies selectable values for parameterizing a query; it cannot import a whole second table's rows into the `orders` dataset for a widget to render.
- D. Global filters constrain existing dataset queries with parameter values; they are not a mechanism for attaching an unrelated table's data to widgets bound to a different dataset.
Sources1
2.Arranging widgets on the canvas
After you have widgets, you arrange them by dragging and resizing with the mouse. A few editing shortcuts make this quicker:
- Hold Command (Mac) or Ctrl (Windows/Linux) and click several widgets to select them together. Moving one moves the whole group. - Right-click a widget and choose Clone. The copy appears below the original, and you edit it like any other widget. - Copy and paste widgets within one dashboard or between dashboards. - Select a widget and press Delete, or right-click it and choose Delete. - Use Command/Ctrl-Z to undo and Command/Ctrl-Shift-Z to redo.
Checkpoint 3 of 6· Check yourself
You want a second chart that matches an existing one except for its x-axis field. What is the quickest route described in the docs?
Cloning a widget puts a copy directly below the original, and you can then change its configuration.
“A clone of your widget appears below the original.”Source: docs.databricks.com
Sources2
3.Text widgets
Charts show the numbers. Text widgets carry the explanation around them, such as headings, notes and links. Drag a text widget onto the canvas and double-click it to edit. The toolbar is the same one used for widget titles, with one addition: Insert images. To edit the raw markdown, open the kebab menu and choose Show markdown.
### Key Highlights
- **Product launches**: Successfully launched 3 new features in March
- **Customer acquisition**: Added 450 new enterprise customers
- **Retention rate**: Maintained 95% customer retention
- **Market expansion**: Entered 2 new geographic marketsCheckpoint 4 of 6· Check yourself
A designer wants the company logo placed inside a bar chart's widget title. What do you tell them?
Titles and descriptions use the same editor as text widgets, except that they cannot contain images.
“You can't insert images into a widget title or description.”Source: docs.databricks.com
Sources3
4.Image widgets and image sources
An image widget shows a single image, such as a logo or diagram, sized to fit the widget. It keeps that fit in draft mode, in published mode, in exports and in embedded dashboards. An image inside a text widget, by contrast, sits within the surrounding text. Both kinds of widget accept the same image sources, and the supported file types are PNG, JPEG/JPG and SVG.
| Source | What to know |
|---|---|
| Direct upload (drag and drop) | Stored in the dashboard definition and limited to 50,000 characters when encoded, about 37 KB. Use one of the other sources for larger files |
| Databricks volume path | Uses the ajax-api files URL shown below. Users must be explicitly granted access to the file |
| Publicly hosted URL | Insert the public URL that points to the image |
Checkpoint 5 of 6· Fill the gap
Complete the URL format for an image stored in a Unity Catalog volume.
https://<databricks-instance>/ajax-api/2.0/fs/ ? /<Volumes/path/to/image>Volume images are served through the /ajax-api/2.0/fs/files/ path. Leave the workspace ID out of the instance name.
Source: docs.databricks.comVolume images have a permission requirement that is easy to miss. Viewers still need explicit access to the image file, even if the dashboard is published with shared data permissions. You grant that access in Catalog Explorer. To get the volume path, open the file in Catalog Explorer and choose Copy Path from its kebab menu.
You can control how an image fills its widget. Fit is the default: it keeps the aspect ratio, never crops, and never shows a scroll bar. Fill stretches the image to the widget's width and adds a scroll bar if it overflows. Alignment positions the image within the widget.
Checkpoint 6 of 6· Match them up
Match each image widget setting to its behaviour.
Tap a term, then the definition that fits it.
Fit is the default and never crops. Fill fills the width and can scroll. Alignment positions the image.
“The image is never cropped, and no scroll bar appears. This is the default.”Source: docs.databricks.com
Sources4
Exam traps
Each one states something that sounds right. Open it to see what is actually true.
1.Widget titles and descriptions use the same editor as text widgets, so they can contain images too.Why is that wrong?
Titles and descriptions support rich text and @ references but not images. Images go in text widgets or image widgets.
Covered in Text widgets
2.If a dashboard is published with shared data permissions, viewers can automatically see images stored in a volume.Why is that wrong?
Whichever data permission mode the dashboard is published with, users must be explicitly granted access to image files stored on Databricks.
Covered in Image widgets and image sources
3.A new visualization widget asks you to choose a dataset before it shows anything.Why is that wrong?
It automatically uses the first dataset on the Data tab and starts as a Bar chart. On a dashboard with several datasets, you need to change the Dataset drop-down yourself.
Covered in Visualization widgets
Practise it for real
Add a text widget and an image widget to a draft dashboard, then check how each one renders.
1.Click New in the sidebar and select Dashboard, then open the Untitled page on the canvas.
Why: Widgets are added to pages on the Canvas, and new dashboards open as drafts.
You should see: An empty draft dashboard with one page named Untitled page.
2.Drag a text widget onto the canvas, double-click it, and paste the Key Highlights markdown from this page.
Why: Text widgets accept markdown as well as the formatting toolbar.
You should see: A heading followed by a bulleted list with bold labels.
3.Open the text widget's kebab menu and click Show markdown.
Why: This shows the raw markdown behind the formatted text.
You should see: The markdown source you pasted, including the ### and ** markers.
4.Drag an Image widget onto the canvas and drop a small PNG file from your device into it.
Why: A direct upload is the simplest image source, but it must stay under about 37 KB once encoded.
You should see: The image appears scaled to fit, centered and uncropped, because Fit is the default.
5.Switch the image widget from Fit to Fill and resize the widget so it is shorter.
Why: Seeing both modes makes the difference clear.
You should see: The image fills the widget's width, and a scroll bar appears when it overflows.
Stuck? Get a nudge
If the upload is rejected, the file is probably too large for a direct upload. Use a volume path or a public URL instead.
Sources
Every claim above is drawn from one of these pages, quoted as it was written on the date shown.
- 1.
“Use the Dataset drop-down to choose the dataset for your visualization.”
↩︎ Visualization widgets“While in draft mode, you can view the full SQL query that a visualization widget runs, including any parameters and filters currently applied.”
↩︎ Visualization widgets“A filter reference reflects the viewer's current selection. A field reference shows a live, aggregated metric.”
↩︎ Visualization widgets“You can't insert images into a widget title or description.”
↩︎ Exam trap 1“By default, the first dataset listed on the Data tab is selected, and the default visualization type is a Bar chart.”
↩︎ Exam trap 3“By default, the first dataset listed on the Data tab is selected, and the default visualization type is a Bar chart.”
↩︎ Prediction“This query is specific to the widget and only returns the data necessary to render that visualization.”
↩︎ Checkpoint“You can't insert images into a widget title or description.”
↩︎ Checkpoint - 2.https://docs.databricks.com/aws/en/dashboards/manageOfficial docs
“When multiple widgets are selected, moving one of them moves the whole selection together as a group.”
↩︎ Arranging widgets on the canvas“To remove a widget, select the widget and press the delete key on your keyboard.”
↩︎ Arranging widgets on the canvas“A clone of your widget appears below the original.”
↩︎ Checkpoint - 3.
“Text widgets let you add formatted text, links, and images to your dashboards.”
↩︎ Text widgets“Drag a text widget onto the canvas and double-click it to start editing.”
↩︎ Text widgets“you must explicitly grant users access to image files stored on Databricks.”
↩︎ Exam trap 2 - 4.
“Image widgets support the same image sources as text widgets: a direct upload, a Databricks volume path, or a publicly hosted URL.”
↩︎ Image widgets and image sources“limited to 50,000 characters when encoded, which is approximately 37 KB”
↩︎ Image widgets and image sources“Image widgets support PNG, JPEG/JPG, and SVG files.”
↩︎ Image widgets and image sources“The image is never cropped, and no scroll bar appears. This is the default.”
↩︎ Checkpoint