Relational data has two basic components:
- Nodes (also called vertices) can represent people, organizations, documents, computers, and so on.
- Ties (also links, relations, edges) can be based on friendship, collaboration, information sharing, resource exchange, etc.
Adjacency and Affiliation
There are two major types of networks that you may need to use in community research: adjacency and affiliation networks. Adjacency networks consist of a single type of nodes and the links between them (example: friendship networks). Affiliation (or two-mode) networks have two types of nodes, often called actors and events. The events, as their name indicates, can be social occasions - but they may also be clubs, teams, companies, schools, topics, preferences, and so on. In affiliation networks there are no direct ties between actors. They are indirectly linked through their affiliation with the same events (they go to the same school, work for the same company, frequent the same pub, drink the same brand of beer, etc.)
Networks can be directed or undirected. Links in undirected networks are by definition reciprocal: if Tom has a family relationship with Jill, then Jill will also have one with Tom. Links in directed networks are not necessarily reciprocated: I go to my physician for health consultations, yet the good doctor will not ask my advice on health issues.
There are a number of important roles that nodes can have within a network:
The Network as a Matrix
There are multiple ways to present and store network data. One of them that you will need to be familiar with is the matrix form. Adjacency networks are presented as square tables with each row and column corresponding to a single network node. Look at the example below: there are four nodes in this small network, and it is represented by a square 4-by-4 matrix. You will notice that row X of column Y equals 1 if there is a link from node X to node Y (and it is 0 if there is no link).