Word Cloud 3D
Signature
sp.wordcloud3d(title, words=None, frequencies=None, points_x=None, points_y=None, edges_i=None, edges_j=None, edges_w=None, *, variant="basic", orientation3d="iso", scene="default", theme="none", zone=None, max_points=None, bg_color="#1a1a2e", width=900, height=500, x_label="", y_label="", z_label="", **kwargs) -> Chart
Aliases: sp.build_wordcloud3d_chart(), sp.word_cloud3d(), sp.wordcloud3d_chart(), sp.tag_cloud3d(), sp.cloud3d().
Description
sp.wordcloud3d() is the 3D twin of sp.wordcloud(): all 8 wordcloud variants share the same 3D form, selected with the same variant keyword. A 2D word cloud’s text packing has no 3D reading, so by default every word instead becomes a column on an outward spiral, height following frequencies. Pass real points_x/points_y (as the 2D chart’s network/context/neuron/cosmos variants compute internally) plus edges_i/edges_j/edges_w and it switches to an honest node-link layout instead, at those exact positions.
Variants
| Variant | Aliases | Uses | 3D form |
|---|---|---|---|
basic | default, spiral, rect, shape, shaped | words, frequencies | One column per word on a frequency spiral. |
image | img, mask, picture, photo, silhouette | words, frequencies | The same spiral as basic; the 2D chart’s mask silhouette has no 3D form of its own. |
labelmap | label_map, datamap, datamapplot, topic_map, scatter_labels | words, frequencies | The same spiral as basic, from the 2D chart’s scatter-label reading. |
network | graph, keywords, co_occurrence, cooccurrence, knowledge_graph | words, frequencies | The same spiral by default; switches to a node-link layout when real positions are supplied. |
bubble | bubbles, packed, circles, packing, pack | words, frequencies | The same spiral as basic, from the 2D chart’s packed-circle reading. |
context | semantic, infranodus, text_network, textnetwork, force, force_directed | words, frequencies | The same spiral by default; switches to a node-link layout when real positions are supplied. |
neuron | neural, brain, synapse, network_glow, nodes | words, frequencies | The same spiral by default; switches to a node-link layout when real positions are supplied. |
cosmos | stars, galaxy, nebula, constellation, space, phyllotaxis | words, frequencies | The same spiral by default; switches to a node-link layout when real positions are supplied. |
Data
words names each column and frequencies sets its height; supplying points_x/points_y (and optionally edges_i/edges_j/edges_w) switches every variant to that node-link layout instead of the spiral.
3D planes
The viewpoint is independent from the variant: orientation3d picks the initial camera plane and applies to every variant; the view always stays draggable.
orientation3d="iso" — Isometric three-quarter view (default).orientation3d="horizontal" — Side-on view, almost at floor level.orientation3d="vertical" — Top-down view from above.orientation3d="front" — Front elevation, nearly flat.3D scenes
scene swaps the environment the columns are drawn in and applies to every variant and every plane.
scene="default" — Each series keeps its own categorical palette color, rendered as shaded 3D bars on a flat grid.scene="terrain" — The data is resampled onto a dense grid and rendered as hundreds of thin colormap-colored columns, forming a continuous mountain-of-bars terrain.scene="tower" — The data is resampled onto a dense field of round skyscraper-like towers instead of boxes, for a city-skyline reading of magnitude.scene="radial" — The data is resampled onto a dense polar grid of rings and wedges radiating from the center, like a 3D radar/sonar sweep of the value field.scene="podium" — The data is resampled onto a dense grid and quantized into stepped terrace levels, like a rice-terrace contour map, to read value bands at a glance.Themes
theme restyles the whole canvas (colour grade and glow) and applies to every variant, scene and plane.
theme="deluxe"theme="prism"theme="aurora"theme="inferno"theme="frost"theme="glow"theme="glass"theme="neon"theme="cosmic"Auto-scaling zone
The 3D zone (floor, walls, axes and camera) scales to the elements: its length, width and height follow the extents of the drawn blocks, a minimum floor depth keeps single rows readable, wide scenes are drawn flatter and the camera frames the whole box. The axis ticks read the real data range. Pass zone=[x, y, z] to force the proportions of the box instead; the longest side is normalised to 1.
Parameters
Returns
Chart object with an .html property and a .show() method.