Adjacency List

The final format commonly used to store network data is known as an adjacency list. The adjacency list takes the dyad-wise representation from an edgelist, and reproduces it in the form of a stem and leaf plot. Table 6.6 presents the adjacency list equivalent of the data represented in Figure 1.1, Table 6.4), and Table 6.5. For each respondent, the adjacency list includes each alter to which they send a tie. There is not a clear norm about whether adjacency lists represent directed or undirected ties, so that information should be included in the metadata. I.e., while the data in Table 6.6 is asymmetrical, that does not necessarily indicate these come from a directed graph.

Table 6.6: Directed, Binary Network in Adjacency List Format.
Resp Alt 1 Alt 2 Alt 3 Alt 4
001 2 3 . .
002 4 . . .
003 2 4 5 .
004 1 5 7 8
005 . . . .
006 5 . . .
007 4 8 . .
008 4 7 . .
009 . . . .

Adjacency lists are convenient, in that they are readily appended onto a flat file data format as additional columns (variables) for each node (observation). They require some manipulation into formats that are more directly used by analytic software packages; most of these packages rely on edgelists or adjacency matrices. However, most SNA software packages have routines available to assist in transformation of data between these standard formats (and even some proprietary formats specific to software packages). For example, in the igraph package for R (Csardi and Nepusz 2006), the “graph” object class that handles igraph’s network data stores an edgelist, but can be read from both an adjacency matrix or adjacency list format. Moreover the intergraph package (Bojanowski 2015) allows for conversion between igraph’s graph objects and the “network” object class used by the statnet package (Handcock et al. 2008). In R, both graph and network objects contain substantially more information than just the tie-level data (e.g., node and graph attributes).