forceNetwork not displaying any edges

问题

I am trying to figure out the EXACT requirements for the node and link dataframes in forceNetwork() because my network is displaying colored movable points without any edges connecting them and I dont know why. As far as I can tell the requirements are:

nodes dataframe must have:

a nodeID column (if nodes arent in the same order as the links) that exactly matches the link source name. If they are in the same order this parameter can be dropped according to my reading of the documentation.

a group variable which specifies the group of the node. Different groups are different colors to differentiate them.

links dataframe must have:

a source column (list of nodes that will connect with targets)

a target column

a value column which specifies the weight of the line to draw between the nodes

> head(links)
  source target value
1  11170      7     1
2   2840      2     1
3  32595      2     1
4  45410      8     1
5  52720     12     1
6  61720      6     1
> head(nodes)
  nodeID group
1  11170     2
2   2840     1
3  32595     2
4  45410     3
5  52720     1
6  61720     2
> head(E(g))
Edge sequence:

[1] 7     -- 11170
[2] 2     -- 2840 
[3] 2     -- 32595
[4] 8     -- 45410
[5] 12    -- 52720
[6] 6     -- 61720
> head(V(g))
Vertex sequence:
[1] "11170" "2840"  "32595" "45410" "52720" "61720"

> typeof(nodes$nodeID[1])
[1] "integer"
> typeof(links$source[1])
[1] "integer"

> dim(links)
[1] 121   3
> dim(nodes)
[1] 135   2

> forceNetwork(Links = links, Nodes = nodes,Source = "source", Target = "target", NodeID = "nodeID",Group = "group", opacity = 0.8, colourScale = "d3.scale.category10()")

回答1:

The source and target vectors in the Links data frame must be numeric, and their values refer to the index of the node in the Nodes data frame which they represent (zero-indexed, unlike R, because it's used by the JavaScript code).

The nodeID vector of the Nodes data frame gives the name (character or numeric) of each node, which is displayed when you hover over a node in the resulting visualization.

Given your example data, I suspect that the data you have in links$source is meant to be the ID of the nodes, and the data you have in links$target is something else. While they're both numeric vectors, the values do not represent the index of the node they refer to in the Nodes data frame.

If you have node names/IDs in your links$source and links$target vectors instead of the index of the nodes in your Nodes data frame, you can fix it like so...

links <- read.table(header = T, text = "
source target value
11170      7     1
2840       2     1
32595      2     1
45410      8     1
52720     12     1
61720      6     1
")

nodes <- read.table(header = T, text = "
nodeID group
11170     2
2840      1
32595     2
45410     3
52720     1
61720     2
")

# build a new Nodes data frame that includes every 
# unique node found in links$source and links$target
nodes_complete <- data.frame(nodeID = unique(c(links$source, links$target)))

# add groups already identified in your original Nodes data frame
nodes_complete$group <- nodes$group[match(nodes_complete$nodeID, nodes$nodeID)]

# convert your links$source and links$target nodes to their index
# in the new Nodes data frame (zero-indexed)
links$source <- match(links$source, nodes_complete$nodeID) - 1
links$target <- match(links$target, nodes_complete$nodeID) - 1

# now the forceNetwork function will run as expected
library(networkD3)
forceNetwork(Links = links, Nodes = nodes_complete, Source = "source", 
             Target = "target", NodeID = "nodeID", Group = "group", 
             opacity = 0.8, 
             colourScale = "d3.scaleOrdinal(d3.schemeCategory10);")

来源：https://stackoverflow.com/questions/30426708/forcenetwork-not-displaying-any-edges

标签

d3.js

networkd3