In this tweet announcing this blog post, @revoandrie visualized his UseR 2015 CRAN network talk with networkD3.

I thought it would be a good opportunity to show a couple more htmlwidgets.

Get All the htmlwidgets

library(igraph)

# devtools::install_github("chrisotphergandrud/networkD3")
library("networkD3")
# devtools::install_github("rich-iannone/DiagrammeR")
#  //rich-iannone.github.io/DiagrammeR
library("DiagrammeR")
# devtools::install_github("dataknowledge/visNetwork")
#  //dataknowledge.github.io/
library("visNetwork")
## Loading required package: magrittr

Code from @revoandrie

This is all code directly from @revoandrie’s gist.

url <- "//rawgit.com/andrie/cran-network-structure/master/pdb/depGraph-CRAN.rds"
datafile <- tempfile(fileext = ".rds")
download.file(url, destfile = datafile, mode = "wb")
gs <- readRDS(datafile)

# Remove all nodes with fewer than 50 edges
deg <- degree(gs, mode = "out")
idx <- names(which(deg > 50))
gn <- induced.subgraph(gs, idx)

# Extract into data frame and plot
gd <- get.data.frame(gn, what = "edges")

networkD3

simpleNetwork(gd, fontSize = 12, height = 600, width = "100%")

I’ll keep the code and customization to hopefully provide a fair comparison.

DiagrammeR

DiagrammeR gives us a couple different options, but the most interactive is vivagraph.

library(htmltools)

# can handle igraph directly
vivagraph(gn,height = 600,width = "100%")

# can even use igraph layouts
vivagraph(gn, layout = layout.auto,height = 600,width = "100%")

visNetwork

visNetwork in my opinion gives the most full-featured interactive network visualization currently.

# use visNetwork

# get a size for our nodes
V(gn)$size <- degree(gn) * 10

gn %>%
  get.data.frame( what = "both" ) %>%
  {
    visNetwork(
      nodes = data.frame(
        id =  .[["vertices"]][,"name"]
        ,size =  .[["vertices"]][["size"]]
      )
      ,edges = .[["edges"]]
      ,height = 600
      ,width = "100%"
    )
  } %>%
  visOptions(highlightNearest = TRUE)