Advanced Usage¶

We run through the advanced functionality of Paxplot using the following synthetic dataset. Note, that Paxplot requires its input be a list of lists or a similar matrix-like format.

data = [
    [0.0, 0.0, 2.0, 0.5],
    [1.0, 1.0, 1.0, 1.0],
    [3.0, 2.0, 0.0, 1.0],
]

Change Number of Ticks¶

By default, Paxplot chooses evenly-spaced ticks between the upper and lower limits of the plotted data. You can change that though!

paxfig = paxplot.pax_parallel(n_axes=4)
paxfig.plot(data)
paxfig.set_even_ticks(
    ax_idx=0,
    n_ticks=15,
)
paxfig.set_even_ticks(
    ax_idx=1,
    n_ticks=16,
    precision=3
)
plt.show()

Custom Ticks¶

Paxplot also gives you the flexibility to set whatever ticks you want, and they can say whatever you want!

paxfig = paxplot.pax_parallel(n_axes=4)
paxfig.plot(data)
paxfig.set_ticks(
    ax_idx=0,
    ticks=[0.0, 1.0, 2.0, 3.0],
    labels=['$my_{heart}$', 'code to', '=', '1612']
)
paxfig.set_ticks(
    ax_idx=2,
    ticks=[0.0, 1.0, 1.5, 2.0],
)
plt.show()

Change Axis Limits¶

By default, Paxplot chooses the axis limits as bounds of the plotted data. You can also change that!

paxfig = paxplot.pax_parallel(n_axes=4)
paxfig.plot(data)
paxfig.set_lim(ax_idx=0, bottom=-1.0, top=3.0)
paxfig.set_lim(ax_idx=2, bottom=1.0, top=3.0)
plt.show()

Axis Inversion¶

Sometimes it’s helpful to invert (flip) an axis. This is particularly helpful when looking at high-dimensional Pareto fronts.

paxfig = paxplot.pax_parallel(n_axes=4)
paxfig.plot(data)
paxfig.invert_axis(ax_idx=0)
paxfig.invert_axis(ax_idx=1)
plt.show()

Adding a Legend¶

It can be nice to plot a legend to identify each line. This works well if you have a few observations.

paxfig = paxplot.pax_parallel(n_axes=4)
paxfig.plot(data)
paxfig.add_legend(labels=['A', 'B', 'C'])
plt.show()

Adding a Colorbar¶

If you have many observations, it is helpful to use a colorbar to distinguish each line. You should also reference the Pandas integration example for another usage of the colorbar.

paxfig = paxplot.pax_parallel(n_axes=4)
paxfig.plot(data)
paxfig.set_label(
    ax_idx=0,
    label='Column A'
)
paxfig.add_colorbar(
    ax_idx=0,
    cmap='viridis',
    colorbar_kwargs={'label': 'Column A'}
)
plt.show()

Accessing Matplotlib Objects¶

Paxplot is an extension of Matplotlib’s subplots wrapper. Paxplot gives you the ability to access the individual Matplotlib axes as well as all the associated functionality using .axes. To demonstrate this, imagine you want to annotate your Paxplot with a label and arrow. That functionality has not been explicitly added to Paxplot, however it does exist for Matplotlib axes. Paxplot still allows us to access that functionality!

paxfig = paxplot.pax_parallel(n_axes=4)
paxfig.plot(data)
paxfig.axes[0].annotate('My Label', (0.3, 0.55))
paxfig.axes[0].arrow(
    x=0.5,
    y=0.52,
    dx=0.0,
    dy=-0.05,
    head_width=0.03,
    head_length=0.02
)
plt.show()

Warning

Access Matplotib axes with caution. Some functions can break your Paxfig object. Using many of these functions will generate a warning from Paxplot.

Getting Started

Examples