Facets and shared scales

Guides

Facets arrange independent Plot values in a uniform grid. Each cell may use different data, marks, labels, and themes. Use scatter_facets() to make grouped scatter panels directly from named DataFrame columns, or build cells explicitly for other marks.

from dataviz import Plot, save_facets

def main() raises:
    var a = Plot().mark_line().encode(
        x=[1.0, 2.0, 3.0], y=[10.0, 15.0, 12.0]
    ).labels(title="North")
    var b = Plot().mark_line().encode(
        x=[1.0, 2.0, 3.0], y=[8.0, 22.0, 18.0]
    ).labels(title="South")
    save_facets(a, b, cols=2, path="facets.svg", shared_y_scale=True)

For a DataFrame, scatter_facets(df, x="spend", y="revenue", facet="region", color="product") returns plots ready for save_facets. facet_order fixes panel order and color_order fixes group colors even when a group is absent from one panel. Both orders must name every observed level. See Grouped DataFrame facets for a complete example.

Use shared_y_scale=True for value-for-value comparison; independent scales are better when each cell’s shape matters. Shared y-scales support point, line, and effect-scatter cells. shared_x_scale=True does the same for x, so a row of panels over one time range ticks and lines up the same way. It takes point, line, area, histogram and effect-scatter cells, widened for any x error bars, and categorical cells that all name the same categories in the same order. A cell with its own scale_x_domain() is refused rather than overwritten. The two flags combine. Every plot must have the same size. The first plot’s labels provide the document-level accessible title for SVG output.

See Facets, Shared facet scale, and facet rendering.

Layers and secondary axes Annotations