Preview Updated 2026-05-10

ParallelCoordinates

Multi-dimensional polylines across parallel Y-axes — the data-science staple for exploratory profiling and outlier spotting.

Basic usage

axes defines each dimension (optional min / max / reversed / format). Each item becomes a polyline, auto-colored by group. Hover focuses one line and dims the rest, perfect for spotting a single instance among many.

背景视口

API

Prop	Type	Default	Description
`axes`	`{ key, label, min?, max?, reversed?, format? }[]`	—	Parallel Y-axis definitions
`items`	`{ label, values: Record<key, number>, group?, color? }[]`	—	Polyline data
`height`	`number`	`280`	Pixel height
`ticks`	`number`	`5`	Tick count per axis
`palette`	`string[]`	viz-1..8	Auto-assigned by group
`highlight`	`number \| null`	`null`	Initial highlighted item index
`ariaLabel`	`string`	—	SR label
`size`	`'sm' \| 'md' \| 'lg'`	`'md'`	Font scale

Payload: { index, item }.

Tip: for an axis where “smaller is better” (e.g. error rate), set reversed: true so “up = good” holds across all axes — comparisons stay visually consistent.