Group Flow Chord Diagram Generator

Free online group flow chord diagram generator. Upload a three-column source,target,value CSV and the tool draws “many sources flowing into a few groups” as a single ring: a ring of small arcs on the left holds every source in one uniform purple, a layer of larger arcs on the right holds your groups — one colour each, matched by a vertical legend — arc width equals each node’s total, and semi-transparent ribbons tinted by the target group span the middle. The default sample renders on load, and you can export high-resolution PNG / SVG.

What a group flow chord diagram is

An ordinary chord diagram places every node equally on one ring, and the ribbons in the middle express how much flows between any pair. A group flow chord diagram introduces a structural division of labour: it assumes your data is bipartite — many small sources on one side, few large groups on the other. The ring is therefore split into two half-circles. The left half is reserved for sources and the right half for groups, each node taking an arc whose angular width equals its own total flow.

The payoff is a very short reading path. To learn what feeds group A, follow its arc inward: every ribbon pointing at it shares one colour, so the composition is visible at a glance. To learn where a source goes, start at its small arc on the left and follow the ribbons branching out of it. Fine white dividers cut each arc into cells proportional to the flow it carries, so the number of cells tells you how many counterparts that node connects to.

What you can do

• Upload a three-column CSV: source,target,value, one row per flow. The source and target name sets must be disjoint — that is the definition of the bipartite structure.
• Automatic bipartite layout: source arcs cluster on the left half, group arcs on the right half, angular widths allocated by total, ribbons tinted by the target group.
• Right-hand group legend: each group gets a colour chip, its name and its total, using exactly the colours of its arc and ribbons.
• White dividers inside group arcs: every group arc is cut into cells by the flows feeding it, so you can count its contributors.
• Default sample equals the template: download the CSV template, upload it back, and you get precisely the chart on screen.
• Light and dark mode: dark mode changes only the background and text contrast, never the chart structure.
• PNG / SVG export: ready for reports, slide decks, posters and dashboards.
• Local processing: everything is parsed and drawn in your browser; nothing is uploaded.

How to use it

1. Open the page and the default sample renders immediately: 14 sources flowing into 8 groups named AAA–HHH.
2. Upload your own CSV, or download the template to get the exact format. Up to 200 rows, ≤24 sources and ≤12 groups.
3. Hover any ribbon to read “source → group : value”. Hover a group arc to see that group’s total.
4. Switch light / dark, then download PNG or SVG.

CSV example:

source,target,value
ppp,AAA,420
ppp,CCC,260
aaa,AAA,380
aaa,BBB,240
ccc,BBB,340
ccc,EEE,200

Where it fits

• Sample classification: which categories sequencing, testing or player samples fall into, and what each category is built from.
• Channel attribution: many channels or landing pages feeding a few conversion sources, revealing the channel mix of each.
• Keyword topic assignment: thousands of keywords assigned to a handful of topic clusters, showing cluster size and source spread.
• Asset and fund grouping: accounts or holdings grouped by asset class or risk band, showing each group’s composition.
• Batch destinations: production batches flowing into a few warehouses or order pools, showing destination concentration.
• Logistics and redistribution: many shipping origins feeding a few regional hubs.
• Cluster review: treat cluster labels as groups and inspect which entities make up each cluster.

Notes

The core idea is bipartite structure plus membership: sources and groups are different kinds of things with very different counts and roles, so they sit on two half-circles instead of being mixed into one. That makes it ideal for many-to-few inflow data. If both sides are comparable in size and you need pairwise flow strength, an ordinary chord diagram fits better; if you need to preserve interval coordinates — which stretch maps to which — use the Circos collinearity chord diagram instead. When the number of groups exceeds the palette capacity, colours are reused cyclically, so merging groups is advisable; the hard limit is 12 groups.

---

Last verified: 2026-09-12

How to read it

Start with the vertical legend on the right: it lists every group with its colour and total, and that is your classification system. Then read the outer layer of large arcs on the right — each arc is one group, longer arcs mean larger totals, and the white dividers cutting the arc into cells mark the individual sources feeding it. Next look at the ring of small purple arcs on the left: each one is a source, and its length is that source’s own total. Finally read the ribbons: colour tells you which group a flow belongs to, width tells you its volume, and the ribbons leaving one small arc add up to that source’s complete destination mix.