Visuals
The library covers bar, line, pie, donut, KPI, gauge, table, pivot table, heat map, tree map, scatter, box plot, funnel, sankey, geospatial maps, word cloud and combo charts.
AutoGraph picks a visual type from the fields you select — useful for exploration, less so for a finished dashboard where you want deliberate choices.
Field wells differ per visual type: X axis, Value, Group/Colour, Small multiples. Small multiples are underused and often better than cramming several series into one chart.
Filters and scope
A filter applies to a single visual, all visuals using this dataset, or all applicable visuals across datasets.
Getting scope wrong is the usual reason a dashboard's numbers do not reconcile. Be deliberate: exploratory filters at single-visual scope, dashboard-wide context filters at dataset scope.
Filters can be exposed as controls — dropdown, list, date picker, slider, relative date — pinned to the top of a sheet so consumers can operate them.
Actions
Filter actions use a selection in one visual to filter others. Navigation actions move between sheets, optionally passing parameter values. URL actions open an external link with field values substituted in.
Actions trigger on select or on menu. Select feels immediate; menu avoids surprising people when a click has several possible meanings.
Sheets and layout
An analysis can have multiple sheets, presented as tabs in the published dashboard.
Layout is either tiled — snapping to a grid, which is the maintainable choice — or free-form, which allows pixel positioning and is worth it only for a polished executive dashboard.
Add a Insights narrative block for auto-generated summaries, and use Q topics if the account has natural-language querying enabled.