Workshop: Relational Event Models - An Introduction to Survival Modelling of Relational Data and Time-Stamped Networks

Date: Wednesday 11 Dec 2019

Time: 12:30pm - 5pm

Location: Oscar Oeser Room 1120 Redmond Barry Building, University of Melbourne

Relational event data are time-stamped events in continuous time that are of a type where one actor does something to another actor. The occurrences of events are traditionally modelled using standard survival analysis techniques but where the past relational history may be taken into account when defining the rates at which events happen. There is a strong connection to network analysis in that when events are aggregated in time, using e.g. sliding time-windows, the patters of pairwise event histories form network graphs. There is also a strong connection to so-called conversational analysis, whereby for example the probability of the next event might be more likely to be actor B responding to actor A than any other act of speech, if the last act of speech was A addressing B.

This half-day training workshop will:

  • Introduce the basic data structure and the fundamentals of relational event models
  • Provide some hands-on worked examples
  • Provide some illustrative examples and an outlook to different implementations.

Preparations: To get the most out of the workshop, please have a look at the seminal article Butts, C. T. (2008). A Relational Event Framework for Social Action, Sociological Methodology, 38: 155-200

Bring a laptop with the program R installed as well as the R-libraries:

  • sna
  • network
  • relevant

Have a look at Tore Opsahl’s tutorials: toreopsahl.com

Program

program