Visualization¶

py3dinterpolations provides built-in visualization for interpolation results using matplotlib (2D slices) and plotly (3D interactive volumes).

3D interactive volume¶

plot_3d_model creates an interactive 3D volume plot using plotly:

from py3dinterpolations import plot_3d_model

fig = plot_3d_model(
    modeler,
    plot_points=True,
    scale_points=10,
    volume_kwargs={"surface_count": 10},
)
fig.show()

Parameters¶

Parameter	Default	Description
`plot_points`	`False`	Overlay training data points on the volume
`scale_points`	`1.0`	Scale factor for point marker sizes
`volume_kwargs`	`None`	Extra kwargs passed to `plotly.graph_objs.Volume`

The returned go.Figure can be customized further using the plotly API, saved as HTML, or displayed inline in Jupyter notebooks.

2D slices¶

plot_2d_model creates a grid of 2D slice plots along a specified axis:

from py3dinterpolations import plot_2d_model

fig = plot_2d_model(
    modeler,
    axis="Z",
    plot_points=True,
    annotate_points=True,
)
fig.savefig("slices.png", dpi=300, bbox_inches="tight")

Each subplot shows one slice along the chosen axis, with optional data point overlay and value annotations.

Parameters¶

Parameter	Default	Description
`axis`	`"Z"`	Axis to slice along: `"X"`, `"Y"`, or `"Z"`
`plot_points`	`False`	Overlay training points on each slice
`annotate_points`	`False`	Annotate point values (requires `plot_points=True`)
`figure_width`	`8`	Figure width in inches

Downsampling comparison¶

plot_downsampling compares original and downsampled data per borehole ID:

from py3dinterpolations.plotting import plot_downsampling
from py3dinterpolations.modelling.preprocessor import Preprocessor

preprocessed = Preprocessor(griddata, downsampling_res=5.0).preprocess()
fig = plot_downsampling(griddata, preprocessed)

This produces a grid of line plots showing the original data (line) and downsampled values (dots) for each borehole, helping assess how downsampling affects the data.

Preprocessing reversal¶

Both plot_2d_model and plot_3d_model automatically reverse preprocessing transformations (normalization, standardization) for display. The plots always show values and coordinates in the original data space.