Dashboards guides
Each guide covers one job, start to finish. They assume you have a service you can edit. If you have never built a dashboard here, the quickstart is a better place to begin, because it covers a little of several of these in one go.
Getting data in
Section titled “Getting data in”Connect a data source — set up a Data Hub connection and get datasets and views showing in the Data panel.
Building
Section titled “Building”Create a dashboard — the creation dialog, and the two choices in it that are permanent.
Add and bind widgets — drag a widget, drag columns onto it, fix the axes and aggregations.
Arrange the canvas — grid, snapping, zoom, multi-select, and sizing charts so their labels fit.
Filter a dashboard — filter widgets on the canvas, and the filter panel readers see in the published view.
Dashboards and AI usage
Section titled “Dashboards and AI usage”Use the AI Assistant while a dashboard is open in the designer to ask for help creating or editing widgets.
Pin data sources before asking
Section titled “Pin data sources before asking”The AI can only use data sources pinned to the current dashboard. You must pin them in the Dashboards module first. A Data Hub connection alone is not enough, and an unpinned dataset or view listed in the Data panel is not available for the AI to use.
- Make sure the datasets or data views you need are available in Data Hub in the same service. Connect a data source covers that setup.
- Open the dashboard in the Dashboards module and choose Data in the designer toolbar.
- Find each dataset or data view you want to use and choose its star-shaped Pin control.
- Confirm those entries appear under Pinned. Pin every source needed for the request, including each source whose fields you want to combine.
- Open AI Assistant and describe the widgets, measures, categories, and layout you want.
Pins are saved with the dashboard. Pinning a source here does not pin it for other dashboards. The star is a data selection for AI, rather than just a favourite for sorting the list.
Ask for a specific result
Section titled “Ask for a specific result”For example, after pinning a monthly sales view, ask: “Using the pinned monthly sales view, add a line chart of total sales by month and a KPI card showing total sales.”
You can also use the AI button beside a pinned source to request a starter dashboard based on that source. This option appears when AI is enabled.
Check the resulting widgets, column mappings, aggregations, and layout before publishing. Add and bind widgets covers how to inspect the bindings, and Publish and version explains how edits reach readers.
If the AI cannot use a source, check that it is under Pinned on this dashboard, then send the request again. If the source is missing from the Data panel entirely, check its availability in Data Hub and that both modules are in the same service. Troubleshooting covers these checks.
Getting it to people
Section titled “Getting it to people”Publish and version — publish a snapshot, see what readers see, and republish after edits.
Schedule a delivery — send a published dashboard as a PDF or image on a recurring schedule.
If something is wrong
Section titled “If something is wrong”Troubleshooting collects the failures that have a known cause, roughly in the order you are likely to run into them.