Arrange related data views around a task while keeping context, comparison, and recovery clear.
A dashboard brings related measures together so people can assess a situation and choose a next step. Define that decision before assembling panels. A collection of unrelated charts can create more work than it removes.
This page adapts general chart guidance to a web composition. Liquidify has no dashboard data engine or chart system; the application owns the data and coordination between views.
Put the most important question first, followed by the comparisons or details needed to act on it. For a support workload view, current queue size may establish the situation, daily arrivals may explain its direction, and a list of unanswered requests may provide the next action.
Use readable headings and visible scope. State the period, active filters, units, and update time where they affect interpretation. A prominent number should not require opening another view to learn what it measures.
Figure in preparation
Organize a data task
Explain a reading order from the current situation to comparison and action.
Keep names and encodings consistent when views share a dataset. Preserve those relationships in a detail view so people can connect it with the summary they selected.
Use comparable ranges when the task requires comparing magnitude across charts. If ranges differ, make that difference visible. Equal-height plots with different bounds can suggest a false comparison. Explain different aggregation periods rather than placing their totals together without context.
Start with the information needed for the current task. Offer a clear path to a longer period, an individual record, or a more detailed view. Keep essential meaning available before interaction.
At narrow widths, prioritize a useful reading order and let panels reflow. Review the available space inside each chart; shrinking an entire desktop dashboard can make ticks and controls unreadable. Preserve access to exact values through an appropriate data presentation.
Distinguish an empty result from an unavailable result. A failed request does not mean a measurement is zero. If previously loaded data remains visible during refresh or failure, identify its age and condition so readers can judge whether to use it.
Keep loading and recovery feedback close to the affected view. A failed secondary chart should not make an otherwise usable dashboard appear complete or entirely unavailable. Offer a meaningful retry path and preserve the current scope where possible.
Use Loading, Empty states, and Notifications for those application-owned sequences.
Gauge can display one value in a range. ProgressIndicator can communicate an operation in progress, and EmptyState can present an explanation and available action. None retrieves data, calculates metrics, or renders a chart series.
Choose layout and selection controls from the component catalog, then implement their relationship to the data explicitly. Review keyboard order, accessible chart descriptions, and the surrounding page as one task.