Scales

Linear Scale

Three linear scales are used here: a scale by index for dots and ticks along the x-axis, a scale for dots and ticks along the y-axis, and a color scale for the dots' fill and stroke style. The associated data is an array of numbers in the range [0, 100]. The first example also demonstrates color derivation using per-channel overrides.

Log Scale

Log scales support color ramps as well.

Diverging Scale

While domains and ranges for quantitative scales are normally uniform, you can specify any number of contiguous regions. For example, for a diverging color scale:

pv.Scale.linear(0, 10, 100).range("red", "yellow", "green")

While the domain is specified explicitly in this example, it could be computed from the data if desired.

Of course, if we combine a log positional encoding with a log color encoding, the ramp is now linear by position:

Ordinal Scale

Like the first example, three scales are used in this chart: an ordinal scale for x, a linear scale for y, and an ordinal scale for color. Now that ordinal scales can be used for categorical color encodings, they may be able to subsume the current pv.Colors.

Notes

Examples:

pv.Scale.linear().range(0, Math.PI) // angle
pv.Scale.linear().range("red", "green") // quantitative colors

Note that for ordinal scales, the domain and range are discrete rather than continuous:

pv.Scale.ordinal().domain("A", "B", "C", "D") // nominal domain
pv.Scale.ordinal().range("red", "green", "blue") // nominal colors