Is there a way to make 4 networks with the same layout?

[fpeti11]

Hello Stefanie, I am quite new in network analysis. I would like to find out if the result that I got is a scale-free network or not? And I would like to measure modularity, is it possible to find out? Could you please help me how to find out these informations? I have a data set that can be divided into 4 categories. I would like to make 4 networks with the very same layout, to be more clear where are the differences. In the tutorial I found only 2 comparable categories, it didn’t work with 4 for me. Do you have a suggestion how to do this? Thank you for your help Peter

[stefpeschel]

Hey Peter,

NetCoMi is actually meant to compare two networks but you can store and reuse the layout so that several networks can be compared visually. Here's the code:

# Construct single network for the first group
net1 <- netConstruct(...)

# Construct single network for the second group
net2 <- netConstruct(...)

# Network analysis for  both groups
netprops1 <- netAnalyze(net1, ...)
netprops2 <- netAnalyze(net2, ...)

# Plot first network
p <- plot(netprops1, ...)

# Plot second network using the first layout
plot(netprops2, layout = p$layout$layout1, ...)

Instead of a single network, you could also construct, analyze, and plot two networks (using the plot arguments sameLayout = TRUE and layoutGroup = "union") and use the joint layout for a third group.

Regarding the scale-free property: You can compute and plot the degree distribution using functions of the igraph package:

net1 <- netConstruct(...)
g <- graph_from_adjacency_matrix(net1$adjaMat1, mode = "undirected", weighted = TRUE, diag = FALSE)
plot(degree_distribution(g), type = "l")

Alternatively, you could also directly take a look at the degree values computed by netAnalyze(), which can be accessed via netprops1$centralities$degree1.

For scale-free networks, the degree distribution should follow a power law.

The network's modularity is computed by netAnalyze() and can be accessed via netprops1$globalProps$modularity1 and is also shown in the summary.

Let me know if you have further questions.

Best, Stefanie