Hiveplotlib

Hiveplotlib is a network visualization library for hive plots, a layout in which nodes are partitioned onto a small number of radial axes and positioned along each axis by a sorting variable. Since positions are determined by the data rather than a layout algorithm, hive plots are reproducible and make structural properties of a graph directly legible.

Below, we visualize Zachary’s Karate Club with both a conventional layout and a hive plot.

import matplotlib.pyplot as plt
import networkx as nx
from hiveplotlib import HivePlot

# sphinx_gallery_thumbnail_number = 2

G = nx.karate_club_graph()

Conventional View: Circular Layout

Zachary’s Karate Club is usually drawn as a circular layout. That view shows every connection, but the relationship between the two factions is hard to read off the resulting alignment of edges.

# color nodes by faction
faction_color = [
    "royalblue" if data["club"] == "Mr. Hi" else "darkorange"
    for _, data in G.nodes(data=True)
]

fig, ax = plt.subplots(figsize=(8, 8))
nx.draw_circular(
    G, with_labels=True, node_color=faction_color, font_color="white", ax=ax
)
plt.show()

Hive Plot View

We pass the NetworkX graph directly to the HivePlot class, partitioning nodes onto axes by their club membership and sorting each axis by node degree (requesting degree as a computed metric at construction time).

Using repeat axes lets us see within-faction edges, while coloring distinguishes within-faction connections (blue/orange) from between-faction ones (gray).

hp = HivePlot(
    graph=G,
    partition_variable="club",
    sorting_variables="degree",
    node_graph_metrics="degree",
    repeat_axes=True,
    non_repeat_edge_kwargs={"color": "darkgray"},
)
hp.update_edges("Mr. Hi", "Mr. Hi", color="royalblue")
hp.update_edges("Officer", "Officer", color="darkorange")

# only two partition groups, so drop one of the two redundant sets of
# between-faction edges
hp.reset_edges(axis_id_1="Mr. Hi_repeat", axis_id_2="Officer")

fig, ax = hp.plot()
plt.show()