Selecting points in one view should light up the same data wherever else it appears on the dashboard

What it is

In a multi-view dashboard, when the user brushes a set of data points in one view, every related view showing the same data should immediately highlight or filter the corresponding content — brushing and linking. The sample boxed in a scatter plot lights up simultaneously in the parallel coordinates, the histogram, and the map; the user's "selection" becomes a shared object across views.

Why it happens

Linking earns its keep by eliminating manual cross-view pairing: without it, the user selects points in view A, must memorise their identities, then hunts for each one in view B — working memory holds three or four, and cross-dimension comparison cost explodes with the dimension count. Linking converts "memorise then hunt" into "select and see," outsourcing working memory to the interface and dropping cross-view comparison cost back to single-view levels. That is the core mechanism of multi-view analysis: each view answers one kind of question well, and linking lets the same data flow between questions without friction.

Where it stops holding

Linking presumes the views share item identity: between aggregate views (already binned) and detail views, linking requires a defined bucket-to-record mapping, and mappings with complicated semantics produce effects the user cannot explain. Synchronisation must be immediate and consistent — partially delayed or inconsistent updates read as malfunction. Nor is more linking always better: full cross-view linking turns one selection into a whole-screen change that users cannot absorb (scope control belongs to the third card in this group).

Applying it

• Enable linked highlighting by default across views of the same data source, with one unified highlight style product-wide.
• When aggregate views participate, state the mapping explicitly ("selecting a bar highlights all records in that month") and note it in the caption.
• Verification: ask users to "select a set of points in the scatter and describe their distribution in the histogram"; completion without manually memorising identities means the linking works.

Related

• Same group: U6.02.2 Linkage relations must be understandable to users · U6.02.3 The scope of linkage needs to be controllable
• Nearby: U6.01.2 Zooming and filtering are two ways to narrow scope · U6.03.1 Applied filters must all be visible
• Search terms: brushing and linking · linked views · multiple coordinated views