Overview

Controls are the dashboard’s filter surface — dropdowns, date pickers, text inputs. Drop one on the canvas, give its variable a name, and that variable is now available for any dataset or chart filter on the dashboard.

Control types

• Dropdown — single select, either from a manual list or from a dataset query.
• Multiselect — multi-select with the same source choices.
• Date — single date.
• Date range — start + end (with presets like "Last 30 days", "Month to date").
• Text input — free-form string.
• Apply button — batches every pending control change. When present, controls don’t auto-fire chart re-queries until the user clicks Apply.
• Refresh button — bumps a revision so every chart re-fetches even without a variable change.

How variables flow

1. Drop a control. It gets a default variable name (e.g. country_1) — rename it via the control’s config tab.
2. Reference the variable in dataset SQL with {{country_1}}, or bind a chart-level filter to it.
3. As the viewer interacts with the control, the value changes, the chart re-queries.

Cascading dropdowns

Two dropdowns where the second filters by the first’s value? Just reference the parent in the child’s options query:

SELECT DISTINCT city FROM customers
WHERE country = {{country}}
ORDER BY city

trend auto-detects the parent dependency and re-runs the child options query whenever the parent changes.

What to read next

• Templating with variables — the SQL side.
• Embed: per-customer attributes — how embed JWTs feed variables.