Boxplot

from __future__ import annotations

from numpy import hstack
from numpy import linspace

from gemseo import configure_logger
from gemseo.datasets.io_dataset import IODataset
from gemseo.post.dataset.boxplot import Boxplot

configure_logger()

<RootLogger root (INFO)>

Build a dataset

inputs = linspace(-1, 1, 100)[:, None]

dataset = IODataset(dataset_name="Foo")
dataset.add_output_variable("y1", inputs**2)
dataset.add_output_variable("y2", hstack((inputs**3, inputs**4)))

other_dataset = IODataset(dataset_name="Bar")
other_dataset.add_output_variable("y1", -(inputs**2))
other_dataset.add_output_variable("y2", hstack((-(inputs**3), -(inputs**4))))

Plot y1 and y2

We can use the Boxplot plot.

plot = Boxplot(dataset)
plot.xlabel = "Variables"
plot.ylabel = "Values"
plot.title = "Standard boxplots"
plot.execute(save=False, show=True)

[<Figure size 640x480 with 1 Axes>]

Plot with centering

We can center the data:

plot = Boxplot(dataset, center=True)
plot.title = "With centering"
plot.execute(save=False, show=True)

[<Figure size 640x480 with 1 Axes>]

Plot with scaling

We can scale the data (normalization with the standard deviation):

plot = Boxplot(dataset, scale=True)
plot.title = "With scaling"
plot.execute(save=False, show=True)

[<Figure size 640x480 with 1 Axes>]

Plot without outliers

We can remove the outliers:

plot = Boxplot(dataset, add_outliers=False)
plot.title = "Without outliers"
plot.execute(save=False, show=True)

[<Figure size 640x480 with 1 Axes>]

Plot with confidence intervals

We can add confidence intervals for the median:

plot = Boxplot(dataset, add_confidence_interval=True)
plot.title = "Confidence intervals"
plot.execute(save=False, show=True)

[<Figure size 640x480 with 1 Axes>]

Plot horizontally

We can use horizontal bars:

plot = Boxplot(dataset, use_vertical_bars=False)
plot.title = "Horizontal bars"
plot.execute(save=False, show=True)

[<Figure size 640x480 with 1 Axes>]

Plot with other datasets

We can add a dataset:

plot = Boxplot(dataset, other_dataset)
plot.title = "Additional dataset"
plot.execute(save=False, show=True)

[<Figure size 640x480 with 1 Axes>]