Choosing a chart
Six peer chart modules, each answering one kind of question and owning its own transition rules.
Each chart module is an independent peer. Its builder adds chart-specific methods to the shared grammar, and the module decides how its marks move. Both endpoints of a transition must use the same chart type.
| Chart | Question it answers | Chart-owned rule |
|---|---|---|
| Bar | How do categories compare? | Bars share a quantitative baseline |
| Line | How does a value change in order? | Keys identify points inside paths; gaps stay honest |
| Area | How does magnitude or composition change? | Explicit baseline and stacked boundaries |
| Point | How do two quantities relate? | Compact Cartesian axes; entering points grow in place |
| Unit | How many countable items are there? | Grid, force, unit-bar and beeswarm layouts; shortest travel |
| Pie | What share of one total does each part take? | Keyed angular motion; direct transitions only |
Every chart can also be partitioned into small multiples
with .repeat().
Import only what you use
import { line } from 'visdelta/line';
const chart = line(rows)
.x('date', { type: 'temporal' })
.y('close');A builder carries its chart module, so transition() loads only the renderer
it needs.
Design warnings
Each chart module can also report advisory design warnings, such as too many stacked columns or a truncated bar baseline. They never change what renders. See Visualization warnings.