Short tutorial on using functions in package “intergraph”

---

“Intergraph” is an R package with coercion routines for netowrk data objects.

• See homepage on http://mbojan.github.io/intergraph.
• Get to package development pages on https://github.com/mbojan/intergraph.

This is a short tutorial showing how to use functions in package “intergraph” using some example network data contained in the package.

Typographical conventions:

• R input code is shown in frames with grey background
• R output is shown in frames with white background with lines preceded with ## (two hash symbols).

For example:

# This is some input code with output below
x <- 2+2
x

## [1] 4

RMarkdown source of this document can be found here. The most up-to-date version is always available as a package vignette.

Example networks

Package intergraph contains four example networks:

• Objects exNetwork and exIgraph contain the same directed network as objects of class “network” and “igraph” respectively.
• Objects exNetwork2 and exIgraph2 contain the same undirected network as objects of class “network” and “igraph” respectively.

All four datasets contain:

• A vertex attribute label with vertex labels. These are letters from a to o.
• An edge attribute label with edge labels. These are pasted letters of the adjecent nodes.
• Network-level attribute layout storing a function that computes the vertex placement for plotting. It is a copy of layout.fruchterman.reingold function from package igraph.

We will use them in the examples below.

To show the data, first load the packages.

library(intergraph)
library(network)

## network: Classes for Relational Data
## Version 1.13.0 created on 2015-08-31.
## copyright (c) 2005, Carter T. Butts, University of California-Irvine
##                     Mark S. Handcock, University of California -- Los Angeles
##                     David R. Hunter, Penn State University
##                     Martina Morris, University of Washington
##                     Skye Bender-deMoll, University of Washington
##  For citation information, type citation("network").
##  Type help("network-package") to get started.

library(igraph)

## 
## Attaching package: 'igraph'
## 
## The following objects are masked from 'package:network':
## 
##     add.edges, add.vertices, %c%, delete.edges, delete.vertices,
##     get.edge.attribute, get.edges, get.vertex.attribute,
##     is.bipartite, is.directed, list.edge.attributes,
##     list.vertex.attributes, %s%, set.edge.attribute,
##     set.vertex.attribute
## 
## The following objects are masked from 'package:stats':
## 
##     decompose, spectrum
## 
## The following object is masked from 'package:base':
## 
##     union

Now, these are the summaries of the “igraph” objects:

summary(exIgraph)

## IGRAPH 3625f1a D--- 15 11 -- 
## + attr: label (v/c), label (e/c)

summary(exIgraph2)

## IGRAPH d2bf475 U--- 15 11 -- 
## + attr: label (v/c), label (e/c)

These are the summaries of the “network” objects:

exNetwork

##  Network attributes:
##   vertices = 15 
##   directed = TRUE 
##   hyper = FALSE 
##   loops = FALSE 
##   multiple = FALSE 
##   bipartite = FALSE 
##   total edges= 11 
##     missing edges= 0 
##     non-missing edges= 11 
## 
##  Vertex attribute names: 
##     label vertex.names 
## 
##  Edge attribute names: 
##     label

exNetwork2

##  Network attributes:
##   vertices = 15 
##   directed = FALSE 
##   hyper = FALSE 
##   loops = FALSE 
##   multiple = FALSE 
##   bipartite = FALSE 
##   total edges= 11 
##     missing edges= 0 
##     non-missing edges= 11 
## 
##  Vertex attribute names: 
##     label vertex.names 
## 
##  Edge attribute names: 
##     label

Networks are shown below using the following code:

layout(matrix(1:4, 2, 2, byrow=TRUE))
op <- par(mar=c(1,1,2,1))
# compute layout
coords <- layout.fruchterman.reingold(exIgraph)
plot(exIgraph, main="exIgraph", layout=coords)
plot(exIgraph2, main="exIgraph2", layout=coords)
plot(exNetwork, main="exNetwork", displaylabels=TRUE, coord=coords)
plot(exNetwork2, main="exNetwork2", displaylabels=TRUE, coord=coords)
par(op)

Functions asNetwork and asIgraph

Conversion of network objects between classes “network” and “igraph” can be performed using functions asNetwork and asIgraph.

network => igraph

Converting “network” objects to “igraph” is done by calling function asIgraph on a “network” object:

# check class of 'exNetwork'
class(exNetwork)

## [1] "network"

# convert to 'igraph'
g <- asIgraph(exNetwork)
# check class of the result
class(g)

## [1] "igraph"

Check if edgelists of the objects are identical

el.g <- get.edgelist(g)
el.n <- as.matrix(exNetwork, "edgelist")
identical( as.numeric(el.g), as.numeric(el.n))

## [1] TRUE

igraph => network

Converting “igraph” objects to “network” is done by calling function asNetwork on an “igraph” object:

net <- asNetwork(exIgraph)

Note the warning because of a “non-standard” network attribute layout, which is a function. Printing “network” objects does not handle non-standard attributes very well. However, all the data and attributes are copied correctly.

Check if edgelists of the objects are identical

el.g2 <- get.edgelist(exIgraph)
el.n2 <- as.matrix(net, "edgelist")
identical( as.numeric(el.g2), as.numeric(el.n2))

## [1] TRUE

Handling attributes

Objects of class “igraph” and “network”, apart from storing actual network data (vertexes and edges), allow for adding attributes of vertexes, edges, and attributes of the network as a whole (called “network attributes” or “graph attributes” in the nomenclatures of packages “network” and “igraph” respectively).

Vertex and edge attributes are used by “igraph” and “network” in a largely similar fashion. However, network-level attributes are used differently. Objects of class “network” use network-level attributes to store various metadata, e.g., network size, whether the network is directed, is bipartite, etc. In “igraph” this information is stored separately.

The above difference affects the way the attributes are copied when we convert “network” and “igraph” objects into one another.

Both functions asNetwork and asIgraph have an additional argument attrmap that is used to specify how vertex, edge, and network attributes are copied. The attrmap argument requires a data frame. Rows of that data frame specify rules of copying/renaming different attributes. The data frame should have the following columns (all of class “character”):

• type: one of “network”, “vertex” or “edge”, whether the rule applies to network, vertex or edge attribute.
• fromslc: name of the which we are converting from
• fromattr: name of the attribute in the object we are converting from
• tocls: name of the class of the object we are converting to
• toattr: name of the attribute in the object we are converting to

The default rules are returned by a function attrmap(), these are:

attrmap()

##      type fromcls  fromattr   tocls       toattr
## 1 network network  directed  igraph         <NA>
## 2 network network bipartite  igraph         <NA>
## 3 network network     loops  igraph         <NA>
## 4 network network     mnext  igraph         <NA>
## 5 network network  multiple  igraph         <NA>
## 6 network network         n  igraph         <NA>
## 7 network network     hyper  igraph         <NA>
## 8  vertex  igraph      name network vertex.names

For example, the last row specifies a rule that when an object of class “igraph” is converted to class “network”, then a vertex attribute name in the “igraph” object will be copied to a vertex attribute called vertex.names in the resulting object of class “network.

If the column toattr contains an NA, that means that the corresponding attribute is not copied. For example, the first row specifies a rule that when an object of class “network” is converted to class “igraph”, then a network attribute directed in the “network” object is not copied to the resulting object of class “igraph”.

Users can customize the rules, or add new ones, by constructing similar data frames and supplying them through argument attrmap to functions asIgraph and asNetwork.

Network objects to/from data frames

Function asDF can be used to convert network object (of class “igraph” or “network”) to a list of two data frames:

l <- asDF(exIgraph)
str(l)

## List of 2
##  $ edges   :'data.frame':    11 obs. of  3 variables:
##   ..$ V1   : num [1:11] 2 3 4 5 6 8 10 11 12 13 ...
##   ..$ V2   : num [1:11] 1 1 1 1 7 9 11 12 13 14 ...
##   ..$ label: chr [1:11] "ba" "ca" "da" "ea" ...
##  $ vertexes:'data.frame':    15 obs. of  2 variables:
##   ..$ intergraph_id: int [1:15] 1 2 3 4 5 6 7 8 9 10 ...
##   ..$ label        : chr [1:15] "a" "b" "c" "d" ...

The resulting list has two components edges and vertexes. The edges component is essentially an edge list containing ego and alter ids in the first two columns. The remaining columns store edge attributes (if any). For our example data it is

l$edges

##    V1 V2 label
## 1   2  1    ba
## 2   3  1    ca
## 3   4  1    da
## 4   5  1    ea
## 5   6  7    fg
## 6   8  9    hi
## 7  10 11    jk
## 8  11 12    kl
## 9  12 13    lm
## 10 13 14    mn
## 11 14 12    nl

The vertexes component contains data on vertexes with vertex id (the same that is used in the first two column of edges) is stored in the first two columns. The remaining columns store vertex attributes (if any). For our example data it is:

l$vertexes

##    intergraph_id label
## 1              1     a
## 2              2     b
## 3              3     c
## 4              4     d
## 5              5     e
## 6              6     f
## 7              7     g
## 8              8     h
## 9              9     i
## 10            10     j
## 11            11     k
## 12            12     l
## 13            13     m
## 14            14     n
## 15            15     o

Functions asNetwork and asIgraph can also be used to create network objects from data frames such as those above. The first argument should be an edge list data frame. Optional argument vertices expectes data frames with vertex data (just like l$vertexes). Additionally we need to specify whether the edges should be interpreted as directed or not through the argument directed.

For example, to create an object of class “network” from the dataframes created above from object exIgraph we can:

z <- asNetwork(l$edges, directed=TRUE, l$vertexes)
z

##  Network attributes:
##   vertices = 15 
##   directed = TRUE 
##   hyper = FALSE 
##   loops = FALSE 
##   multiple = FALSE 
##   bipartite = FALSE 
##   total edges= 11 
##     missing edges= 0 
##     non-missing edges= 11 
## 
##  Vertex attribute names: 
##     label vertex.names 
## 
##  Edge attribute names: 
##     label

This is actually what basically happens when we call asNetwork(exIgraph)

---

Appendix

sessionInfo()

## R version 3.2.2 (2015-08-14)
## Platform: x86_64-pc-linux-gnu (64-bit)
## Running under: Ubuntu 14.04.3 LTS
## 
## locale:
##  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
##  [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
##  [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
##  [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
##  [9] LC_ADDRESS=C               LC_TELEPHONE=C            
## [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       
## 
## attached base packages:
## [1] stats     graphics  grDevices utils     datasets  methods   base     
## 
## other attached packages:
## [1] igraph_1.1.0     network_1.13.0   knitr_1.11       intergraph_2.0-2
## 
## loaded via a namespace (and not attached):
##  [1] magrittr_1.5    formatR_1.2.1   tools_3.2.2     htmltools_0.2.6
##  [5] yaml_2.1.13     stringi_0.5-5   rmarkdown_0.8   stringr_1.0.0  
##  [9] digest_0.6.8    pkgconfig_2.0.0 evaluate_0.8