Legends and visual encodings

An encoding maps a data column to a visual channel. Numeric x and y control position. color, color_categories, size, and labels add continuous color, categorical color, point size, and per-point text.

from dataviz import Plot, save

def main() raises:
    var plot = (
        Plot()
        .mark_point()
        .encode(
            x=[1.0, 2.0, 3.0, 4.0],
            y=[8.0, 5.0, 9.0, 6.0],
            color_categories=["East", "West", "East", "West"],
        )
        .labels(title="Revenue by Region")
    )
    save(plot, "encoding.svg")

Legends are derived from encodings; flat Theme.mark_color does not need one. For layered charts, call .series_name() on each layer to create one legend row per named layer. Keep the number of categories small enough to distinguish, and use shape as well as color when hue alone is insufficient.

See Color by category, Layer legend, and Plot.encode.