Spectral Ring Chord Diagram Generator

A free online spectral ring chord diagram generator. Upload a three-column source,target,value CSV and turn a pairwise relationship matrix into a colored ring: nodes are laid clockwise from the largest total to the smallest, and each arc takes its color from one continuous spectral ramp (red, orange, yellow, green, teal, blue), so hue itself is the ranking — the biggest node is red and the smallest leans blue. Every relationship becomes a uniform thin teal line running from the midpoint of one node’s arc segment to the other, bending in toward the centre and weaving into a mesh when dense, with a translucent pink ring anchoring the middle. The demo chart loads instantly and exports to high-resolution PNG / SVG.

What a spectral ring chord diagram is

A classic chord diagram encodes flow volume as ribbon width, which forces you to compare thicknesses. This variant takes a different route: it gives order to color and relationships to thin lines. Arc length is still allocated by each node’s total, but colors are not drawn from a repeating palette — they are sampled along one continuous spectral ramp by node rank. Rank 1 lands on the red end, the last rank on the blue end, and everything in between falls into orange, yellow, green and teal. You therefore never need a legend: redder means bigger, bluer means smaller.

The second difference is the shape of the links. Instead of width-mapped ribbons, every relationship is a uniform 2px line drawn from the midpoint of each node’s arc segment, connected by a soft curve that bends toward the centre. The payoff shows up in dense networks: wide ribbons paint over each other into a solid smear, while thin lines weave a mesh whose routes stay traceable. You can follow a single line all the way from A to B, and you can see at a glance which nodes sit at the hubs.

Three details sit on the ring. Inside each arc, white dividers split it by each relationship’s value, so the divider count equals that node’s number of connections. A small black dot marks the centre of every arc. And node names sit outside the ring, rotated along the tangent, so they never eat into the drawing area.

What you can do

• Upload three columns: source,target,value, one row per relationship, with value required to be positive.
• Automatic symmetric merge: A→B and B→A are combined into a single two-way relationship with their values added, so no connection is drawn twice.
• Ranked clockwise layout: nodes are ordered by total from largest to smallest, with the biggest fixed at twelve o’clock.
• Continuous spectral sampling: colors are interpolated along the ramp by node index, so hue reads as rank and never repeats.
• Uniform thin links: all relationships share the same 2px teal stroke, forming a legible network texture at high density.
• White dividers inside arcs: each arc is split by its individual relationship values — the cell count is that node’s connection count.
• Pink centre ring: a translucent anchor for the eye that carries no data.
• Sample equals template: download the CSV template, upload it back, and you get exactly the chart on screen.
• Light / dark mode: dark mode only changes the background and text contrast, never the structure.
• PNG / SVG export: ready for reports, slides, posters and dashboards.
• Local processing: everything is parsed and drawn in your browser, never uploaded.

How to use it

1. Open the page and the default sample is already drawn: 28 country nodes forming a dense network after 119 raw flow rows are merged into two-way links.
2. Upload a CSV, or grab the format with “Download CSV template”. Up to 200 rows and ≤60 nodes.
3. Hover any thin line to see “node ↔ node : value”. Hover an arc to see that node’s total.
4. Switch light or dark, then download PNG or SVG.

CSV example:

source,target,value
USA,CHN,540
CHN,USA,520
USA,CAN,400
CAN,USA,420
DEU,NLD,200
NLD,DEU,220

The first two rows, USA,CHN,540 and CHN,USA,520, merge into one USA ↔ CHN relationship worth 1060.

Where it fits

• Co-occurrence and co-words: keyword, tag or product co-occurrence matrices, to spot topic clusters and hub terms.
• Trade and logistics: two-way volume between countries or regions, to see which nodes dominate and which sit at the edge.
• Citations and collaboration: paper or patent citations and co-authorship networks, to find core institutions and clusters.
• Calls and messages: exchange counts between numbers or accounts, to surface unusual hubs and ring structures.
• Ecological and biological networks: species interactions, protein interactions, gene co-expression.
• Internal operations: ticket, approval or email traffic between departments and teams.
• Migration: city-to-city in and out flows, to read the two-way intensity profile.

Notes

The core idea is color carries order, thin lines carry relationships: arc length still follows totals, but color is sampled continuously along a spectrum, so hue becomes a ranking scale and one layer of legend reading disappears; links skip width mapping to preserve traceable paths in a dense network. That makes it best suited to moderately sized, densely connected matrices (roughly 10 to 60 nodes). If you need to compare the strength of individual relationships precisely and width differences matter more to you, a classic chord diagram is the better fit; if your data splits naturally into a many-and-small side and a few-and-large side, use the group flow chord diagram instead.

---

Last verified: 2026-09-17

How to read this chart

Start with color: sweep clockwise from the reddest arc, and the hue slides gradually toward blue — that ramp is the ranking scale, so the reddest node has the largest total and the bluest the smallest, with no legend needed. Then look at arc length: each arc’s angular width is proportional to its total, and the red arc is usually the longest. Next, the white dividers inside an arc split it by each relationship, so more dividers mean more connections; the black dot at each arc’s centre marks the node. Finally, the inside: one teal line is one two-way relationship, its ends landing on segment midpoints of two different arcs and bending toward the centre before opening out — stacked together, the lines reveal hub nodes (many lines converge) and peripheral ones (only one or two). The pink ring in the middle is purely a visual anchor and encodes nothing.