Moving from critique to construction, the quote implies that good visualizations begin with a task: detect anomalies, compare groups, track change, or locate relationships. Shneiderman’s broader work in human-computer interaction argues for aligning interactive systems with what users are trying to accomplish, rather than what designers want to showcase.
Consequently, the best design choices tend to look practical: axes that support accurate comparison, labels that reduce mental effort, and encodings chosen for the specific question at hand. When the user’s task drives the form, insight becomes the natural outcome rather than an accident. [...]
