About WebShapes

WebShapes is a web platform that allows researchers and practitioners to create various network shapes using different methods and parameters. Users can upload their own networks and download the shape information — including boundary points, fitted parameters, and 3D figures — for further analysis.

The underlying methodology is described in:

Jin, Shengmin, and Reza Zafarani. "Representing Networks with 3D Shapes." IEEE International Conference on Data Mining (ICDM), 2018.  PDF →

Jin, Shengmin, and Reza Zafarani. "The Spectral Zoo of Networks: Embedding and Visualizing Networks with Spectral Moments." ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2020.  PDF →

The Spectral (m₂, m₃, m₄) embedding method computes the 2nd, 3rd, and 4th spectral moments of the random walk transition matrix P = D⁻¹A. These moments are interpretable: m₂ relates to degree distribution, m₃ to triangle density, and m₄ to squares in the network. All values lie in [0, 1], forming a compact unit-cube embedding space.