sankey plot

Sankey plot / Sankey diagram / Sankey chart / Alluvial Flow Diagram (Highly Customizable and Publication‑quality)

Basic usage - links

links={'a1','A',1.2;'a2','A',1;'a1','B',.6;'a3','A',1; 'a3','C',0.5;
    'b1','B',.4; 'b2','B',1;'b3','B',1; 'c1','C',1;
    'c2','C',1;  'c3','C',1;'A','AA',2; 'A','BB',1.2;
    'B','BB',1.5; 'B','AA',1.5; 'C','BB',2.3; 'C','AA',1.2};

% Create a Sankey diagram object
% 创建桑基图对象
SK = SSankey(links(:,1), links(:,2), links(:,3));

% Start drawing
% 开始绘图
SK.draw()

Basic usage - adjMat

% Define inter-layer adjacency matrices
% 定义层间邻接矩阵
A12 = [1,2,1; 1,2,3; 2,0,1];
A23 = [1,4; 2,1; 0,3];
A34 = [1,5; 2,3];

% Assemble global block matrix (main diagonal = zero, super-diagonal = A12, A23, A34)
% 组装全局分块矩阵（主对角线为零，上对角线为 A12, A23, A34）
adjMat = mergeAdjMat({A12, A23, A34});

SK = SSankey([],[],[], 'AdjMat',adjMat);
SK.draw()

Feature under testing

A12 = [1,2,1; 1,2,3; 2,0,1];
A23 = [1,4; 2,1; 0,3];
A34 = [1,5; 2,3];
adjMat = mergeAdjMat({A12, A23, A34});

SK = SSankey([],[],[], NodeList = compose('C%d',1:size(adjMat, 1)), AdjMat = adjMat).draw();

% Add nodes and links using the + operator (使用 + 运算符添加节点和链接)
% add link: obj + [source, target, value]
% add node: obj + {'Name', layer}
SK + {'add1', 1} + {'add2'} + {} + {'layerNode5', 5} + [13, 14, 5] + [10, 14, 3] + [12, 14, 5];

This MATLAB toolkit facilitates the creation of highly customizable and publication‑quality Sankey diagrams. It supports both link‑cell and adjacency‑matrix inputs, and offers a range of rendering modes ('left', 'right', 'interp', 'map', 'simple'), four vertical alignment options ('up', 'down', 'center', 'justify'), and multiple link‑interpolation methods ('pchip', 'linear', 'bezier', 'makima', 'spline'). Interactive features include callback support for link clicks.

Key layout parameters—such as block gaps, label positions, value‑label offsets, block coordinates, and link resolution—are fully customizable. The package also provides utilities for bidirectional chord diagrams and vertical Sankey layouts, along with over 20 usage demos and a replication of a *Nature* journal figure.

The compressed package contains numerous usage examples.