This repo constitutes an R package, and contains data and tools for the triadic analysis of affiliation networks.

Description

The paper Triadic analysis of affiliation networks makes a case for adopting a batch of triad-based tools for the study of (bipartite) affiliation networks. Most of the tools used therein are included in this package, which is built mostly on the igraph package. No new classes have been defined yet, but several back-end functions are written in C++ without reference to the igraph library. Any suggestions would be welcome.

Installation

The package is not yet on CRAN; it can be installed using the devtools package:

devtools::install_github("corybrunson/bitriad")

Functionality

The package implements several tools from the paper, most importantly

  • triad_census(), which surveys the triads of an affiliation network and returns the census in a specified scheme; and
  • triad_closure(), which surveys the wedges of an affiliation network and returns either global or local proportions of wedges that are closed.

The parameters for these functions, in particular the census schemes and the definitions of wedge and closure, are thoroughly documented in help(triad_census) and help(triad_closure). Both functions pass to their corresponding functions in igraph when the input graph is not an affiliation network.

Datasets

Empirical affiliation networks from the following sources are included as datasets:

Documentation

The vignette southern_women outlines an analysis of women_clique and women_group using the censuses, some clustering coefficients, and other tools.

The full documentation is built into a bitriad website using pkgdown.