Skip to content

canvod-viz

Purpose

The canvod-viz package provides 2D and 3D hemispheric visualization for GNSS-T grids, SNR data, and VOD results. It wraps matplotlib (publication-quality polar plots) and plotly (interactive 3D hemispheres) behind a unified API.

Components

  •   HemisphereVisualizer

    Unified entry point combining both 2D and 3D backends. Swap between publication and interactive modes with one method call.

  •   HemisphereVisualizer2D

    Matplotlib polar projection plots. Patch-based rendering of grid cells. Publication-quality output at configurable DPI.

  •   HemisphereVisualizer3D

    Plotly interactive 3D hemispheres. Pan, zoom, rotate in the browser. One-call HTML export for sharing.

  •   Tissot Indicatrix

    add_tissot_indicatrix — overlay angular distortion circles on 2D plots to evaluate grid cell shape fidelity.

Usage

from canvod.grids import create_hemigrid
from canvod.viz import HemisphereVisualizer

grid = create_hemigrid("equal_area", angular_resolution=10.0)
viz  = HemisphereVisualizer(grid)

# Publication-quality 2D
fig_2d, ax_2d = viz.plot_2d(data=vod_data, title="VOD Distribution")

# Interactive 3D
fig_3d = viz.plot_3d(data=vod_data, title="Interactive VOD")
fig_3d.show()

from canvod.viz import visualize_grid, visualize_grid_3d, add_tissot_indicatrix

fig, ax = visualize_grid(grid, data=vod_data, cmap="viridis")
add_tissot_indicatrix(ax, grid, n_sample=5)

fig_3d = visualize_grid_3d(grid, data=vod_data)

viz = HemisphereVisualizer(grid)
viz.set_style(create_publication_style())

fig, ax = viz.create_publication_figure(
    data=vod_data,
    title="VOD Distribution",
    save_path="figure_3.png",
    dpi=600,
)

viz.set_style(create_interactive_style(dark_mode=True))

fig = viz.create_interactive_explorer(
    data=vod_data,
    dark_mode=True,
    save_html="explorer.html",
)

(fig_2d, ax_2d), fig_3d = viz.create_comparison_plot(data=vod_data)

Styling

Style factory Use case
create_publication_style() Print-ready figures, configurable DPI, white background
create_interactive_style() Browser-ready Plotly, dark mode option
create_rse_style() Remote Sensing of Environment journal guidelines

Both return a PlotStyle / PolarPlotStyle object passed to viz.set_style().

RSE Journal Style

Publication-quality styling matching Remote Sensing of Environment guidelines: Arial/Helvetica fonts, 300 DPI, inward ticks, colorblind-friendly palette (Wong, 2011).

from canvod.viz import rse_context

with rse_context():
    fig, ax = plt.subplots()
    ax.plot(x, y)

from canvod.viz import rse_style

@rse_style
def make_figure():
    fig, ax = plt.subplots()
    ax.plot(x, y)
    return fig, [ax]

from canvod.viz import apply_rse_style
apply_rse_style()  # modifies plt.rcParams globally

Colorscale

Cross-framework colormap conversion between matplotlib, Plotly, and palettable:

from canvod.viz import Colorscale

cs = Colorscale.from_matplotlib("viridis", n_colors=64)
plotly_cs = cs.to_plotly()     # [(0.0, 'rgb(...)'), ...]
mpl_cmap  = cs.to_matplotlib() # LinearSegmentedColormap

cs2 = Colorscale.from_colors(["#0072B2", "#D55E00", "#009E73"])

Interactive Grid Exploration

The demo/20_grid_exploration.py marimo notebook provides interactive exploration of per-cell VOD data on the hemispheric grid:

  • 3D view — Plotly Scatter3d with click-to-select cells (anywidget)
  • 2D view — Canvas polar projection with hover tooltips (anywidget)
  • Reactive stats — selected cells display in a table with mean, std, count, coverage
  • Timeseries — per-cell VOD timeseries plotted for selected cells
uv run marimo edit demo/20_grid_exploration.py

Dependencies

Optional backends

  • matplotlib — required for HemisphereVisualizer2D and all 2D functions
  • plotly — required for HemisphereVisualizer3D and all 3D functions
  • anywidget — required for interactive hemisphere selectors in marimo notebooks
  • canvod-grids — always required for grid geometry

Neither backend is a hard dependency of canvod-viz itself; import errors are raised only when the corresponding visualizer is instantiated.