Legend Sizing
Legend Sizing
Use this when: You need to shrink a categorical legend’s own swatches and row spacing.
Usage
from dataviz import Plot, save
from dataviz.core.theme import Theme
def main() raises:
var x: List[Float64] = [
1.0,
2.0,
3.0,
4.0,
5.0,
6.0,
7.0,
8.0,
9.0,
10.0,
11.0,
12.0,
]
var y: List[Float64] = [
12.0,
18.0,
15.0,
22.0,
19.0,
25.0,
21.0,
28.0,
17.0,
24.0,
20.0,
27.0,
]
var team: List[String] = [
"Alpha",
"Bravo",
"Charlie",
"Delta",
"Echo",
"Foxtrot",
"Golf",
"Hotel",
"India",
"Juliett",
"Kilo",
"Lima",
]
var plot = (
Plot()
.mark_point()
.encode(x=x, y=y, color_categories=team)
.labels(
title="Readings by Team",
subtitle="A compact legend for many categories",
)
.theme(Theme(legend_swatch_size=8, legend_row_gap=4))
)
save(plot, "docs/src/examples/out_legend_sizing.svg")Related
Related recipes: Smoothed Line · Continuous Legend Size
Relevant API: Theme · OutputFormat