Social Network Analysis
This post presents an example of social network analysis with R using package igraph.
The data to analyze is Twitter text data of @RDataMining used in the example of Text Mining, and it can be downloaded as file "termDocMatrix.rdata" at the Data webpage. Putting it in a general scenario of social networks, the terms can be taken as people and the tweets as groups on LinkedIn, and the term-document matrix can then be taken as the group membership of people. We will build a network of terms based on their co-occurrence in the same tweets, which is similar with a network of people based on their group memberships.
At first, a term-document matrix, termDocMatrix, is loaded into R. After that, it is transformed into a term-term adjacency matrix, based on which a graph is built. Then we plot the graph to show the relationship between frequent terms, and also make the graph more readable by setting colors, font sizes and transparency of vertices and edges.
Note that the above termDocMatrix is a standard matrix, instead of a term-document matrix under the framework of text mining. To try the code with your own term-document matrix built with the tm package, you need to run the code below before going to the next step.
Transform Data into an Adjacency Matrix
Build a Graph
Now we have built a term-term adjacency matrix, where the rows and columns represents terms, and every entry is the number of co-occurrences of two terms. Next we can build a graph with graph.adjacency() from package igraph.
Plot the Graph
A different layout can be generated with the first line of code below. The second line produces an interactive plot, which allows us to manually rearrange the layout. Details about other layout options can be obtained by running ?igraph::layout in R.
Make it Look Better
Next, we will set the label size of vertices based on their degrees, to make important terms stand out. Similarly, we also set the width and transparency of edges based on their weights. This is useful in applications where graphs are crowded with many vertices and edges. In the code below, the vertices and edges are accessed with V() and E(). Function rgb(red, green, blue, alpha) defines a color, with an alpha transparency. We plot the graph in the same layout as the above figure.
More examples on social network analysis with R and other data mining techniques can be found in my book "R and Data Mining: Examples and Case Studies", which is downloadable as a .PDF file at the link.