manynet: an R package for multimodal network analysis and more

Presenters: James Hollway and Henrique Sposito

This workshop introduces manynet (previously migraph), an R package for the analysis of many networks. The package aims to make learning, teaching, and applying network analysis easier across a wider range of types of networks. The package includes measures of network, tie, and node centrality, cohesion, closure, and diversity, and routines to count motifs and identify roles and community membership. It also includes CUG and QAP tests and models, blockmodelling, and network diffusion and learning models. The package operates with many common formats, classes, and types of network. There are read and write functions for several common external formats, including edgelists, Pajek, and UCINET, as well as coercion to and from common R classes, such as matrices and dataframes, as well as igraph, network, and tidygraph objects. All functions in the package work for all classes and a wide range of types of networks: e.g. (un)directed, (un)weighted, (un)signed, as well as one-mode, two-mode and three-mode networks, using sensible defaults and appropriate normalisations. The goal of this workshop is to introduce participants to the package through tutorials using multimodal network data. Familiarity with R and RStudio is recommended. Participants can bring their own research problems and data and, depending on the number of participants, remaining time can be used to discuss them.