plot_all_metric_pareto_front
site.SiteSolutionSet.plot_all_metric_pareto_front(
height=4,
show_points=True,
theme='whitegrid',
maxx=None,
maxy=None,
cols=3,
**kwargs,
)Plot Pareto fronts for all pairs of solution metrics.
This method generates a grid of subplots, each showing the Pareto front for a pairwise combination of performance metrics. It provides a comprehensive view of trade-offs between all available objectives.
Parameters
| Name | Type | Description | Default |
|---|---|---|---|
| height | float | Height (in inches) allocated to each subplot. | 4 |
| show_points | bool | If True, all solutions are plotted as points in addition to the Pareto front in each subplot. | True |
| theme | str | Visual theme passed to the underlying plotting function. | "whitegrid" |
| maxx | bool or None | If True, x-axis metrics are treated as maximisation objectives when computing Pareto fronts. If False, they are minimised. If None, the direction is inferred per metric. | None |
| maxy | bool or None | If True, y-axis metrics are treated as maximisation objectives when computing Pareto fronts. If False, they are minimised. If None, the direction is inferred per metric. | None |
| cols | int | Number of columns in the subplot grid. | 3 |
| **kwargs | Additional keyword arguments passed to spv.pareto_plot. |
{} |
Returns
| Name | Type | Description |
|---|---|---|
| matplotlib.figure.Figure | The matplotlib Figure containing all Pareto front subplots. |
Notes
The method constructs all pairwise combinations of the following metrics: - “weighted_average” - “unweighted_average” - “90th_percentile” - “max” - “proportion_within_coverage_threshold” (included only if available)
Each subplot visualises the Pareto front for a pair of metrics using the spv.pareto_plot function.
Subplots are arranged in a grid with a specified number of columns, and rows are determined automatically.
Any unused subplot axes (if the grid is larger than required) are removed from the figure.
The figure is closed before returning to prevent duplicate display in some environments (e.g., Jupyter notebooks).