Graph layouts using cluster information. The function calculates a
grouped layout based on the node attribute in igraph or tbl_graph
object, while respecting the original edge attributes. If you performed
community detection based on the network, these layouts should not be
used as this layout just explicitly places nodes belonging to the same
category together. The layout_cluster_wise uses igraph::merge_coords
internally to place nodes using the dla algorithm. Reference to
Group Attributes Layout in Cytoscape and component_wise and
layout_components for the disconnected graphs in igraph.
## Make an example graph and assign a random category
library(clusterlayouts)
library(igraph)
test <- igraph::random.graph.game(n=100, p=0.01)
V(test)$category <- as.factor(sample(1:4, 100, replace=TRUE))
lyt <- clusterlayouts::layout_cluster_panel(test, "category")
plot(test, layout=as.matrix(lyt[,1:2]),
vertex.color=as.numeric(V(test)$category),
vertex.size=5, vertex.label=NA)
## Can change per cluster layout
test <- igraph::random.graph.game(n=500, p=0.001)
panel <- c(2,3,5,6,6,3,2,2)
V(test)$category <- as.factor(sample(1:sum(panel), 500, replace=TRUE))
lyt <- clusterlayouts::layout_cluster_panel(test, "category", per_row=panel, per_layout="drl")
#> Overriding nrow option
plot(test, layout=as.matrix(lyt[,1:2]),
vertex.color=as.numeric(V(test)$category),
vertex.size=3, vertex.label=NA)
## Based on the column
test <- igraph::random.graph.game(n=500, p=0.001)
panel <- c(2,3,5,6,6,3,2,2)
V(test)$category <- as.factor(sample(1:sum(panel), 500, replace=TRUE))
lyt <- clusterlayouts::layout_cluster_panel_col(test, "category", per_col=panel)
#> Overriding ncol option
plot(test, layout=as.matrix(lyt[,1:2]),
vertex.color=as.numeric(V(test)$category),
vertex.size=3, vertex.label=NA)
## Can change widths per column
test <- igraph::random.graph.game(n=1000, p=0.001)
panel <- c(2,4,3,1,5,3,6,7,2,4)
widths <- c(1,2,3,4,5,12,3,2,1,2)
V(test)$category <- as.factor(sample(1:sum(panel), 1000, replace=TRUE))
lyt <- layout_cluster_panel_col(test, "category",
per_col=panel, widths=widths)
#> Overriding ncol option
plot(test, layout=as.matrix(lyt[,1:2]),
vertex.color=as.numeric(V(test)$category),
vertex.size=1, vertex.label=NA)