NetworkAdvanced

Hive Plot

A perceptually uniform network layout using radial axes to position nodes by attribute — making network structure readable, reproducible, and comparable.

// 01 — The chart

What it looks like

Example — Software dependency network3 axes · by connectivity
Axis 1: SourcesAxis 2: LibrariesAxis 3: UtilitiesHigh-degree hub

A hive plot with three axes showing software modules positioned by connectivity. The highlighted edge traces a key dependency from a source hub to a utility node.

// 02 — Definition

What is a hive plot?

A hive plot is a network visualization method invented by Martin Krzywinski in 2012 to address the fundamental problem of force-directed layouts: they look different every time you run them. In a hive plot, nodes are placed along linear axes radiating from a central point, with their position on each axis determined by a measurable node attribute (such as degree, age, or cluster membership).

Typically a hive plot uses 2–4 axes. Nodes are assigned to an axis based on one categorical attribute (like node type or community), then positioned along that axis by a quantitative attribute (like connectivity or centrality). Edges are drawn as curved lines connecting nodes on different axes.

The key advantage is perceptual uniformity: the same data always produces the same layout, making hive plots ideal for comparing networks, spotting structural differences, and communicating reproducible findings. Unlike force-directed graphs, there’s no randomness — the layout is entirely determined by data.

Origin: The hive plot was introduced by Martin Krzywinski at the BC Cancer Research Centre in 2012, published in Briefings in Bioinformatics. It was designed specifically to create rational, reproducible network visualizations for genomics and systems biology research.

// 03 — Anatomy

Parts of a hive plot

ABCDE
A — Central origin: The point from which all axes radiate, representing the conceptual center of the network
B — Radial axis: A linear axis extending from the center; each axis represents a node category or grouping
C — Node position: Placement along the axis is determined by a quantitative attribute (e.g., degree, centrality)
D — Curved edge: A Bézier curve connecting nodes on different axes, representing a relationship in the network
E — Node size: Optional encoding of an additional attribute through node radius, adding a third data dimension

// 04 — Usage

When to use it — and when not to

✓Use a hive plot when…
  • You need a reproducible, deterministic network layout that doesn’t change between runs
  • Nodes have meaningful categorical attributes that map naturally to 2–4 axes
  • You want to compare two networks side by side using the same layout rules
  • Your network has known node types (e.g., proteins, genes, metabolites) that define axes
  • You need to present network findings in a publication or report where reproducibility matters
  • You want to reveal connection patterns between specific node categories
×Avoid a hive plot when…
  • Your nodes lack meaningful categorical attributes to define axes
  • You want an exploratory, free-form layout — use a force-directed graph instead
  • The audience is unfamiliar with network visualization — hive plots have a learning curve
  • You have more than 4 node categories — too many axes become confusing
  • Your network is very small (<10 nodes) — a simple node-link diagram is clearer
  • You need to show geographic or spatial positions of nodes

// 05 — Reading guide

How to read a hive plot

Follow these steps whenever you encounter a hive plot.

1

Identify the axes and their meaning

Each axis represents a category of nodes. Read the axis labels to understand what grouping is being used — node type, community, functional role, or another categorical attribute.

2

Understand the position metric

Nodes are placed along their axis based on a quantitative attribute. Nodes closer to the center typically have lower values (e.g., fewer connections), while those further out have higher values. Check the legend or caption for the metric.

3

Read the edges between axes

Curved lines connecting nodes on different axes represent relationships. Dense bundles of edges between two axes indicate strong inter-group connectivity.

4

Look for edge density patterns

Where edges cluster between axes reveals the dominant connection patterns. Sparse zones between axes mean those groups interact less frequently.

5

Compare with other hive plots

Because hive plots are deterministic, you can place two side by side and spot structural differences — a missing bundle, a shift in density, or new high-degree hubs.

// 06 — Common mistakes

Mistakes to watch out for

Using meaningless axis assignments

The power of a hive plot comes from intentional axis assignment. Randomly assigning nodes to axes produces no insight. Each axis should represent a meaningful category that you want to compare.

Too many axes

More than 4 axes makes the plot confusing and hard to interpret. The angles between axes become too narrow, edges overlap unpredictably, and the visual advantage is lost. Stick to 2–4 axes.

Ignoring the position metric

If node position along the axis is arbitrary, you lose half the information the plot can convey. Always map position to a meaningful quantitative variable (degree, centrality, expression level).

Not labeling axes clearly

Without clear axis labels and position metric descriptions, readers can’t interpret the plot. Always annotate what each axis represents and how nodes are ordered along it.

Overloading with too many edges

Dense networks create edge bundles that obscure all structure. Filter edges by weight threshold, or show only the top connections to keep the visualization readable.

// 07 — Real-world examples

Where you’ll see hive plots used

01

Genomics: Protein interaction networks

Researchers use hive plots to visualize protein–protein interaction networks in model organisms. Axes represent functional categories (kinases, receptors, transcription factors), and position encodes connectivity. Side-by-side hive plots compare healthy vs. diseased interaction networks.

Bioinformatics
02

Software engineering: Module dependencies

Software architects map codebase dependencies using hive plots. Source modules, libraries, and utilities each get an axis, with position determined by import count. This reveals which modules are over-coupled and where dependency refactoring is needed.

Software Analysis
03

Social networks: Community interactions

Network scientists use hive plots to compare social network structures across platforms or time periods. Axes represent detected communities, and the deterministic layout allows precise before/after comparisons when studying structural changes.

Network Science

// 08 — At a glance

Quick reference

Also known asHive panel, rational network layout
Invented byMartin Krzywinski, 2012
Best forReproducible, attribute-driven network visualization
Data typesNetwork data with categorical + quantitative node attributes
Recommended axes2 – 4 axes for clarity
Key propertyDeterministic. Same data always produces same layout.
Common toolsD3.js, Cytoscape, HiveR (R package), custom implementations
Common mistakesMeaningless axes, too many axes, unlabeled positions, edge overload

// 09 — Variations

Types of hive plots

The basic hive plot can be adapted with different axis configurations and edge encodings.

Two-axis hive plot

The simplest form with just two axes, useful for visualizing connections between two distinct node types.

Four-axis hive plot

Uses four axes for networks with four node categories. More expressive but requires careful labeling to remain readable.

Paired hive plot (comparison)

Two hive plots side by side using identical axis rules, designed for comparing network structure across conditions or time.

Weighted hive plot

Edge thickness encodes relationship strength or frequency, adding a quantitative dimension to the connection pattern.

// 10 — FAQs

Frequently asked questions

What is a hive plot?+

A hive plot is a network visualization method invented by Martin Krzywinski in 2012 to address the fundamental problem of force-directed layouts: they look different every time you run them. In a hive plot, nodes are placed along linear axes radiating from a central point, with their position on each axis determined by a measurable node attribute (such as degree, age, or cluster membership).

When should you use a hive plot?+

Use a hive plot when you need a reproducible, deterministic network layout that doesn’t change between runs. It also works well when nodes have meaningful categorical attributes that map naturally to 2–4 axes, and when you want to compare two networks side by side using the same layout rules.

When should you avoid a hive plot?+

Avoid a hive plot when your nodes lack meaningful categorical attributes to define axes. It is also a poor fit when you want an exploratory, free-form layout — use a force-directed graph instead, or when the audience is unfamiliar with network visualization — hive plots have a learning curve.

Is a hive plot suitable for dashboards?+

Yes — a hive plot can work well in dashboards as long as the panel is large enough for readers to perceive the encoded values, has a clear title, and includes the legend or axis labels needed to interpret it.

What category of chart is a hive plot?+

Hive Plot belongs to the Network family of charts. Charts in that family are designed to answer the same kind of question, so they often work as alternatives when one doesn't quite fit your data.

How do you read a hive plot?+

Start with the axis labels and legend, then look at the overall shape before zooming into individual marks. Compare prominent features against the rest of the data, and verify any conclusion against the underlying numbers when precision matters.