Home | Intro to SNA | WN-SNA

WN-SNA

Cohesion, centrality and other ‘whole network’ measures and metrics.

Krack-Sociom dataset

This dataset was compiled with a sociometric survey of 21 executives working in a hi-tech (for the time) startup. Respondents were give a full roster of (21) names and asked to nominate who they went to for advice (Advice relation), and who they considered a friend (Friendship relation). We also know who each respondent reported to (Report relation). The dataset ...

Read More »

Whole networks – Density

Random graphs The top-down WN-SNA perspective on network data has generated a lot of work with random graph modelling. This modelling uses computer simulations to explore ideas about the emergence of networks. For mathematicians, graphs are abstract models of network data, they are constructed entities defined as a set of nodes and specifications of di-ties between pairs; the basic SNA ...

Read More »
Globe and geodesics

Paths and Geodesic distances

Network software can trace all the paths that connect any two nodes. There are many possible paths. SNA focuses on the shortest paths which it calls ‘geodesic paths’ or simply ‘geodesics’. (A geodesic is the shortest path on a curved surface. Thus the geodesic paths on a conventional atlas – Mercator projection – will appear as arcs.) What is a ...

Read More »

Node centrality: Degree based measures

I think of (node) centrality measures as giving an ‘inside-out’ perspective on the network within which the node is embedded. Calling it centrality implies a ‘top-down’ perspective and a centralized organization of network diagrams. Remember that the layout algorithm for network diagrams worked by bringing the most connected nodes to the centre. We see this with the operation of Netdraw’s ...

Read More »

Node centrality: Path-based measures

NEEDS SCREENSHOTS The closeness centrality scores are derived from matrices of geodesic paths we generated yesterday. They are path-based measures rather than degree-based measures. Node scores for closeness are made by summing along the row of geodesics for each node. You can use Univariate Stats on the – geo output to see this. Closeness centrality measures Note that the Multiple ...

Read More »