Cluster ego-centered networks by a grouping factor

The idea of clustered graphs is to reduce the complexity of an ego-centered network graph by visualizing alters in clusters defined by a categorical variable (Lerner et al. 2008). clustered_graphs() calculates group sizes, inter and intra group tie densities and returns these informations in a list of igraph objects.

Usage

clustered_graphs(object, ..., clust.groups)

# S3 method for class 'list'
clustered_graphs(object, aaties, clust.groups, ...)

# S3 method for class 'egor'
clustered_graphs(object, clust.groups, ...)

# S3 method for class 'data.frame'
clustered_graphs(object, aaties, clust.groups, egoID = ".egoID", ...)

Arguments

object: An egor object.
...: arguments to be passed to methods
clust.groups: A character naming the factor variable defining the groups.
aaties: data.frame/ list containing alter-alter relations as a 'global edge list' or as a list of 'edge lists'. (not needed if object is an egor object).
egoID: Character. Name of the variable identifying egos (default: "egoID").

Value

clustered_graphs returns a list of graph objects representing the clustered ego-centered network data;

References

Brandes, U., Lerner, J., Lubbers, M. J., McCarty, C., & Molina, J. L. (2008). Visual Statistics for Collections of Clustered Graphs. 2008 IEEE Pacific Visualization Symposium, 47-54.

Examples

data("egor32")

# Simplify networks to clustered graphs, stored as igraph objects
graphs <- clustered_graphs(egor32, "country") 

# Visualise
par(mfrow = c(2,3))
vis_clustered_graphs(
  graphs[1:5]
)

par(mfrow = c(1,1))