I'm using the iGraph package in R to layout a network graph, and I would like to group the vertex coordinates based on attribute values.
Similar to the answered question How to make grouped layout in igraph?, my question differs in that the nodes needn't be grouped by a community membership that was derived from a community detection algorithm.
Rather, I want to layout with groups based on attribute values that are known in advance for each vertex.
For example, if each vertex has an attribute "Master.Org", and there are ~10 to ~20 distinct values for Master.Org, then how can I layout the graph such that all vertices within the same Master.Org are grouped ?
Thanks!
Additional Detail
In fact, two separate attributes provide nested levels of grouping.
My goal is to layout a graph object such that the "Master.Org" and "Org.Of" values are grouped together in their XY coordinates on the graph.
For example, each node will belong to an "Org.Of". And there can be multiple "Org.Of" values within the "Master.Org".
Thoughts ?
Thanks!
While this question is rather old, it is a reasonable question and deserves an answer.
No data was provided so I will generate an arbitrary example.
library(igraph)
set.seed(1234)
G = erdos.renyi.game(20, 0.25)
V(G)$Group1 = sample(3,20, replace=TRUE)
plot(G, vertex.color=rainbow(3, alpha=0.4)[V(G)$Group1])
Without doing anything, the Group is ignored.
Now, we need to create a layout that will plot nodes
in the same group close together. We can do this by creating
a graph with the same nodes, but with additional links between
nodes in the same group. The within-group links will be given
a high weight and the original links will be given a small weight.
This will cluster nodes in the same group. We then apply the
layout to plotting the original graph, without the extra links.
They were just to get a good layout.
G_Grouped = G
E(G_Grouped)$weight = 1
## Add edges with high weight between all nodes in the same group
for(i in unique(V(G)$Group1)) {
GroupV = which(V(G)$Group1 == i)
G_Grouped = add_edges(G_Grouped, combn(GroupV, 2), attr=list(weight=5))
}
## Now create a layout based on G_Grouped
set.seed(567)
LO = layout_with_fr(G_Grouped)
## Use the layout to plot the original graph
plot(G, vertex.color=rainbow(3, alpha=0.4)[V(G)$Group1], layout=LO)
If you want to go beyond this to have multiple levels of grouping, just add additional links with appropriate weights to connect the subgroups too.
Related
It's my first time working with igraph - so please excuse my basic questions. I'm doing a stakeholder analysis and have different actors on different administrative scales.
I'm trying to plot my network in a hierarchical way, because I want to show the interaction between administrative boundaries.
My vertices have the attributes "name", "Label" and "Scale".
I want to plot the graph with Scale as the layers.
edges <- data.table::fread("Master_edges.csv")
nodes <- data.table::fread("Master_nodes.csv")
V(g1_2)$name <- nodes$Label
V(g1_2)$Scale <-nodes$Scale
l <- layout_with_sugiyama(layers=V(g1_2)$Scale) #this is wrong, how do I define the layers?
plot(g1_2, layout=l, vertex.label=V(g1_2), vertex.color=NA)
Also I have different types of edges - how can I plot the different edge-types in different plots?
I would like to create temporal networks in R but the only resources I've found works with FR or KK graphs. However, my primary graph that I would like to base the layout from is a DRL layout. How could I code this in R to keep the layouts?
Thank you
Added:
Code:
drl <- layout.drl(netfull, options=list(simmer.attraction=0))
plot(netfull, edge.arrow.size=2, vertex.size=.5, vertex.label.cex=.3, vertex.label.dist=.1, vertex.lable.degree=pi, layout=drl)
plot(net7, edge.arrow.size=2, vertex.size=.5,vertex.label.cex=.3, vertex.label.dist=.1, vertex.lable.degree=pi, layout=drl)
You can just explicitly compute your layout before plotting and then use the layout argument when you want to plot. DRL is one of the standard options provided by igraph.
library(igraph)
## create test graph
set.seed(1234)
g = erdos.renyi.game(15, 0.2, type = "gnp")
## Create a reusable layout for the graph
LO = layout_with_drl(g)
## plot using the layout
plot(g, layout=LO)
Edit
Based on the discussion in the comments, I have a different understanding of the question. I think that the question is this: Given a graph g and a subgraph g2 print both g and g2 with the corresponding nodes in the same place. This extra response addresses that.
Start with the example above to create the graph g and the layout LO.
Now we want to take a subgraph and print it with the corresponding nodes in the same place. I will use as an example the graph that we get by removing nodes 2, 9, and 15.
If we simply remove those nodes, the new graph will have 12 nodes and they will have node IDs 1-12. In order to preserve the original numbering, we need to save the node IDs as labels.
V(g)$label = 1:15
Now let's create the subgraph by removing nodes 2,9 and 15.
g2 = induced_subgraph(g, V(g)[-c(2,9,15)])
We want to reuse the layout LO, but LO has the positions for all 15 original nodes. We want to select only the part for the remaining nodes in g2.
LO2 = LO[-c(2,9,15),]
Now we are ready to plot the original graph and the reduced graph so that the nodes line up.
par(mfrow=c(1,2), mar=c(2,1,2,1))
plot(g, layout=LO, frame=TRUE)
plot(g2, layout=LO2, frame=TRUE)
I created an igraph with a community membership identified:
fc <- fastgreedy.community(graph)
colors <- rainbow(max(membership(fc)))
This provided me the clusters that each of the nodes belong to.
Now when I plot this:
plot(graph,vertex.color=colors[membership(fc)],
layout=layout.kamada.kawai)
it doesn't provide a layout where it exclusively separates each group of nodes based on the membership. Does anyone know a different layout that can provide this? All this is doing is taking the layout: kamada.kawai and coloring in the memberships rather than restructuring the layout so that it is organized by membership.
Hope this question makes sense. Thanks!
You have to calculate the Kamada-Kawai layout with an artificial weight vector that assigns a high weight to edges within clusters and a low weight to edges that cross cluster boundaries:
> graph <- grg.game(100, 0.2) # example graph
> cl <- fastgreedy.community(graph)
> weights <- ifelse(crossing(cl, graph), 1, 100)
> layout <- layout_with_kk(graph, weights=weights)
> plot(graph, layout=layout)
The trick here is the ifelse(crossing(cl, graph), 1, 100) part -- crossing(cl, graph) takes a clustering and the graph that the clustering belongs to, and returns a Boolean vector that defines for each edge whether the edge is crossing cluster boundaries or not. The ifelse() call then simply replaces TRUE (i.e. edge crossing boundaries) in this vector with 1 and FALSE (i.e. edge stays within the cluster) with 0.
Using the R Kohonen package, I have obtained a "codes" plot which shows the codebook vectors.
I would like to ask, shouldn't the codebook vectors of neighbouring nodes be similar? Why are the top 2 nodes on the left so different?
Is there a way to organise it in a meaningful organisation such as this image below? Source from here. Where the countries of high poverty are clustered at the bottom.
library("kohonen")
data("wines")
wines.sc <- scale(wines)
set.seed(7)
wine.som <- som(data = wines.sc, grid = somgrid(5, 4, "hexagonal"))
# types of plots
plot(wine.som, type="codes", main = "Wine data")
Map 1 is the average vector result for each node. The top 2 nodes that you highlighted are very similar.
Map 2 is a kind of similarity index between the nodes.
If you want to obtain such kind of map using the map 1 result you may have to develop your own plotting function with the following parameters:
Pick up the most relevant nodes or the most different ones (manually or automatically). Then, you have to attribute a color to each of these nodes.
Give a color the the neigbours nodes using the average distance between the center of each node from the selected nodes. Shorter distance = close color, higher distance = fading color.
To sum up, that's a lot of work for nearly nothing. Map 1 is better and contains a lot of informations. Map 2 is nice looking...
I want to create a graph plot with each community of nodes been covered by some background color, similar to the graph by the following code
karate <- graph.famous("Zachary")
wc <- walktrap.community(karate)
modularity(wc)
membership(wc)
plot(wc, karate)
But different from this approach, I want to: (1) group the nodes by myself, instead of resulting from community detection algorithm. I achieved this by wc$membership <- some_vector; (2) plot such graph possibly with overlap between communities, then how can I assign one node to multiple communities?
Plot the graph itself instead of the community structure, and use an appropriately constructed mark.groups argument to plot() to tell igraph which groups should be enclosed by the same group. The following quote is from the manual page of plot.igraph:
mark.groups: A list of vertex id vectors. It is interpreted as a set of vertex groups. Each vertex group is highlighted, by plotting a colored smoothed polygon around and "under" it.