Introducing HyperNetX: A Python Library for Complex Network Data Analysis (306535)Cliff Joslyn, PNNL
*Brenda Praggastis, PNNL
Emilie Purvine, PNNL
Keywords: Network Science, Hypergraphs, Topological Data Analysis, Python libraries
Hypergraph models of complex high-dimensional data provide faithful representations of arbitrary, multi-way relationships, such as those found in collaboration networks or biological interactions. Once solely the purview of graph theory, network measures and analysis methods, such as distance, connectivity, and clustering, are now available for "hypernetwork" models. Moreover, hypergraphs allow for higher-order methods from computational topology that go beyond the study of two-way interactions. HyperNetX(HNX) is an open source Python library, inspired by NetworkX and recently released by PNNL, which provides a platform for exploratory data analysis and visualization of hypergraphs. This talk will introduce hypergraphs and the metrics they provide. We will demonstrate multiple use cases, illustrating a variety of approaches for modeling complex data using hypergraphs and discuss what these models tell us about the data.