Data Structures and Variables for NetSense

There are three types of data:

• Event data: indicating when a communication event occurred and between whom (and other information)

• Dyadic data: For a give time period, frequencies of interaction between two nodes. Created by collapsing event data over a time period and over node pairs.

• Node data: For a given time period, counts of people i communicated with (degree), along with activity levels

For the dyadic and node data the notation references nodes i and j, and time period t.

e_ijt = 1. An edge exists if in time period t (e.g., week 1) i initiated a communication event with j.

If e_ijt = e_jit = 1, the tie is mutual in that week.

We defined a new tie, e*ijt = 1 as one where eijt= 1 and ejit’ = 1 where t’ ≤ t. That is one edge has occurred in that week or a prior week, and now the other edge occurs in week t.

We define an active tie in a given week, aijt = 1, if e*ijt’ = 1 for t’ ≤ t and eijt = 1 or ejit= 1