Beyond a few lines, tracking collapses

What it is

Overlaying many series creates crossings, occlusion, and similar appearances that increase identity-tracking and legend-matching cost, producing the familiar spaghetti plot. “A few” in the title is not a fixed ceiling. Trackability depends on task, crossing density, line separation, sample count, chart size, visual ability, and interaction. An overview of the envelope or outlying trajectories may tolerate many lines; accurately following one series or comparing a specified pair often breaks down earlier.

Why it happens

Readers must maintain series identity along a spatial path and rebind it at crossings or close approaches using positional continuity, colour, stroke pattern, and labels. More series often create more opportunities for crossings and occlusion, but geometry rather than count alone determines difficulty. A remote legend adds repeated travel between name and path. Highlighting, direct labels, filtering, and facets reduce the candidate set or place identity near the path; they do not repair an invalid scale or missing semantics.

Studying it

With realistic data shapes, manipulate series count, crossing density, colour and stroke discriminability, direct labels, faceting, and interactive focus. Separate named-path tracing, endpoint identification, pairwise comparison, anomaly detection, and ensemble-summary tasks. Measure accuracy, response time, pointer or gaze paths, and switches to the wrong line after crossings. Include delivered sizes, zoom, colour-vision variation, low vision, and keyboard use; ratings of clutter alone cannot establish task performance. A practical threshold should follow acceptable error consequences, not a universal line count.

Where it stops holding

A spaghetti plot is not useless for every task. Background trajectories can convey distribution, variability, and outliers if highlighted and summary layers do not erase individual evidence. Aggregation lowers tracing demand but may hide minority groups, extremes, or high-consequence cases. Facets improve local clarity yet make cross-panel comparison harder when scales or spacing diverge. Interactive highlighting needs keyboard operation, a visible state, and alternatives for touch and exported images.

Applying it

• Let the task determine whether to preserve individual trajectories, foreground selected targets, combine summaries with context, or split into shared-scale small multiples. Do not impose a fixed line-count ban.
• Put series names at distinguishable endpoints or near paths and offer search, filtering, hover, and keyboard focus. Change stroke weight, lightness, or labels as well as colour when highlighting.
• Provide a series list, key-value table, trend summary, and point-wise navigation so screen-reader users can select a series and inspect its status, range, and anomalies.
• Test named-path, endpoint, pairwise, anomaly, and ensemble tasks. If critical tasks produce frequent line switches after crossings, reduce simultaneous candidates, facet, or choose another comparison structure.

Related

• Same group: U2.03.1 Line charts suit trends over ordered variables · U2.03.2 Never connect a categorical axis with lines
• Adjacent: U2.15.1 Small multiples replace overlay with repeated like-for-like panels · U1.12.3 Highlight only the class currently in focus
• Search terms: spaghetti plot · line tracing · small multiples · direct labeling · interactive highlighting