Laplacian Library

A Repository of Graph Laplacians, originally developed for Graph Signal Processing.

Note: The synthetic datasets do not yet have dedicated edge files as of 2026.

Dataset	Vertices	Signals	Edges	Signals
StanfardBunny	35947	x,y,z	104288
Horse	48485	x,y,z	145449
SyntheticHawaii	37	long, lat	98	length, admitt, delay
SyntheticTexas	2000	long, lat		length, admitt, delay
SyntheticEast	70000	long, lat		delay
SyntheticEastWest	78000	long, lat	100923	length, admitt, delay
SyntheticUSA	82000	long, lat	104123	length, admitt, delay
WECC	240		451	length, admitt, delay

Subfolder Structure

Each dataset subfolder should adhere to the following structure:

DatasetName/
├── edges.txt      # Sparse edge list
├── signals.txt    # Vertex features
├── lap.mat        # Graph Laplacian
└── README.md      # Metadata

Edge Data (edges.txt)

A space-separated text file with a header line containing graph edges and features.

• Dimensions: M x (2 + K) matrix (M edges, K features).
• Columns: source, target, followed by K feature columns.
• Indices: 0-based integers.
• Notes: Supports multiple edge weights (e.g., resistance, reactance).

Vertex Signals (signals.txt)

A space-separated text file representing an N x D matrix of vertex features.

• Rows: N vertices, sorted by index (0 to N-1).
• Columns: D signal dimensions/features.

Graph Laplacian (lap.mat)

A binary MATLAB Level 5 MAT-file containing the N x N graph Laplacian matrix.

• Variable: Must be named A.
• Format: Compressed Sparse Column (CSC).

Metadata (README.md)

A markdown file containing dataset metadata. Each file should include:

• Description (str)
• Number of Vertices (int)
• Number of Edges (int)