Network Dynamic Temporal Visualization (ndtv): A Package of R Tools for Visualizing Dynamic Networks

Day: Tuesday 11:30 am – 2:30 pm

Workshop Length: 1 session (3 hours)

Attendance Limit: <20

This workshop will provide an introduction to creating animated movies of dynamic networks using the two R

based packages ndtv (http://cran.r-project.org/web/packages/ndtv/) and networkDynamic. These tools can be used for both empirical and simulated network data. We will demonstrate both, with some well-known data sets from the social network literature, and some simulations from the statnet package tergm

Topics will include exporting movies and HTML5 web animations, the tradeoffs of various layout algorithms, considerations about how to slice and aggregate time in networks, various modes of representing relationships in time, and how to effectively use a range of graphical

properties to represent dynamic attributes of nodes and edges (color, shape, size, transparency, speed, and

annotation). We will also discuss some common challenges, such as adjustments when working with networks with disconnected components, and how to determine if a network has appropriate size and density to create an animation. Some non-animation techniques such as relationship timelines will be explained as well.

The workshop will also cover the basic functionality of the networkDynamic package (http://cran.r-project.org/web/packages/networkDynamic), including how to import, transform and extract relational data with timing information from various data structures (matrices, spell lists, toggles, etc). We will discuss advantages of various temporal models and representations (continuous vs discrete time, etc). Attaching and manipulating dynamic vertex and edge attributes will be demonstrated.

Prerequisites: Familiarity with R. Previous experience with the statnet packages (ergm, tergm, network and networkDynamic) is helpful but not required.

 Submitting Instructor: Skye Bender-deMoll

Institution: statnet

Email: skyebend@skyeome.net