In jr-packages/jrAutomate: R Automate Exercises

library("edgebundleR")

Overview

This package creates a hierarchical edge bundle plots; similar to the D3 example. The main function in the package is edgebundle(). It takes in a variety of inputs:

• an igraph object
• a symmetric matrix, e.g. a correlation matrix or (regularised) precision matrix
• a JSON file structured with name and imports as the keys

You can install this (large) package using

install.packages("edgebundleR")

Examples

library(edgebundleR)
library(MASS)
data(mtcars, package = "datasets")
sig = kronecker(diag(3), matrix(2, 5, 5)) + 3 * diag(15)
X = MASS::mvrnorm(n = 100, mu = rep(0, 15), Sigma = sig)
colnames(X) = paste(rep(c("A.A", "B.B", "C.C"), each = 5), 1:5, sep = "")
edgebundle(cor(X), cutoff = 0.2, tension = 0.8, fontsize = 14)

It can also be used for displaying correlation matrices. We simply calculate the correlation and set any correlations between $-0.7$ and $0.7$ to be zero

cor_mt = cor(mtcars)
cor_mt[cor_mt < 0.7 & cor_mt > -0.7] = 0

Then plot as usual

edgebundle(cor_mt)

Further information

• github

jr-packages/jrAutomate documentation built on Dec. 14, 2019, 6:35 p.m.