voronoi

Partition the plane into Voronoi cells around a set of points and color each cell. This is the standard way to render a profile-reduced field over irregular support points without interpolating to a grid — each support point becomes one colored cell, so the rendering is faithful to the reduced data.

Based on: a custom Jarvis-PLOT renderer built on scipy.spatial.Voronoi (no direct Matplotlib equivalent).

Axes: rectangular (ax) and ternary (axtri).

Coordinates

Key	Required	Meaning
x	yes	Generator point x
y	yes	Generator point y
z	no	Value per cell. With z, cells are colored by colormap; without z, all selected cells get one flat color

Style

Color mode (z present):

Key	Type	Purpose
cmap	string	Colormap
vmin / vmax	float	Color scale limits
edgecolor	color	Cell edge color ("none" to hide)
linewidth	float	Cell edge width
alpha	0–1	Opacity
zorder	number	Draw order

Flat-fill mode (z absent):

Key	Type	Purpose
where	bool array	Fill only cells where the mask is true
facecolor	color	Flat fill color
edgecolor	color	Cell edge color
radius / extent	—	Bound the cells (advanced)

Example — profile-likelihood map

- name: prof
  data:
    - source: df
      transform:
        - profile:
            bin: 150
            objective: max
            grid_points: rect
            coordinates:
              x: {expr: mC1, name: xx, lim: [0, 1.4]}
              y: {expr: mN1, name: yy, lim: [0, 1.4]}
              z: {expr: LogL, name: z0}
  share_data: prof_grid
  axes: ax
  method: voronoi
  coordinates:
    x: {expr: xx}
    y: {expr: yy}
    z: {expr: z0}
  style:
    cmap: jarvis_rainbow2_r
    vmin: -50
    vmax: 0
  colorbar: axc

Notes

• voronoi requires scipy.
• For a smoothed version, interpolate the reduced points with make_interp_2d and render with pcolormesh. The profile_2d figure type lets you switch between the two with interp: true|false.
• For boundary-only / hatched styling, see voronoif.

---

See also: Plot Methods index · voronoif · pcolormesh