Workshops

The conference will feature two days of workshops on beginning and advanced topics in network analysis along two parallel tracks.

We recommend that attendees pre-install the software listed in italics before arriving at the workshops.

• R is required for almost all workshops,
• RStudio is highly recommended, and
• statnet is frequently required, and
• xergm 1.5 is required for the advanced ERGM workshops.

Please follow the instructions on the statnet installation page for the first three. Installing statnet in that way will also install ergm, sna, and network.

Please install the latest xergm 1.5 from Philip Leifeld's page (source and Windows and Mac binaries) for the advanced ERGM workshops.

Links to data files for individual presentations can be downloaded by clicking on the specific workshop below.

The workshop lineup is as follows:

*A download location for the latest version of xergm will be sent out shortly before the workshops.

**qgraph, RColorBrewer, extrafont, and ggplot2 also recommended.

June 17, 9:00–12:30pm

Introduction to Network Theory and Methods - Michael Heaney

This workshop introduces the questions, concepts, theories, methods, and important findings of social network analysis for scholars that have no prior training in this field. It begins by discussing the logic of relational thinking and motivations for the study of networks, including the identification of appropriate research questions and fertile areas for research. The collection of network data using various methodologies is considered, such as surveys, interviews, experiments, archival analysis, and ethnography. It discusses the dilemmas of boundary specification, the trade-offs of ego-network and whole-network approaches, and the challenges of asking network questions in surveys. It presents definitions of key objects (e.g., links, nodes, modes, components, paths, triads), one-mode versus multi-modal analysis, and network statistics (e.g., degree, degree distribution, indegree, outdegree, density, centrality, centralization). It surveys the major theories in network analysis (e.g., balance, embeddedness, resource dependence, brokerage, status, homophily, multiplexity, structural holes, preferential attachment, and small world theory), as well as recent findings from studies of political behavior, legislatures, parties, law, advocacy, trade, and international conflict and cooperation.

Basic ERGMs - Skyler Cranmer

This workshop introduces the student to statistical inference on network data via the exponential random graph model (ERGM). We will begin by explaining the dangers of mistreating network data and then introduce the basic ERGM. We will discuss the specification, estimation, and interpretation of the model. We will also show and discuss several practical examples using the ERGM package in R.

June 17, 1:30–5:00pm

Intro to Network Analysis in R - Lorien Jasny

This workshop session will serve as a basic introduction to the importation, manipulation, and descriptive analysis of social network data within the R/statnet platform. Topics covered will include: basics of R syntax; an overview of basic R functions and data types; importation of network data into R; network data manipulation; management of metadata for complex networks; visualization of network data; calculation of network descriptives (e.g., centrality scores, graph-level indices); and use of classical network analytic techniques (e.g., blockmodeling). No prior experience with R or statnet is assumed, but attendees should have familiarity with the basic concepts of descriptive network analysis.

Longitudinal Models [TERGM/SAOM] - Philip Leifeld

The workshop on modeling dynamic networks focuses mainly on the Temporal Exponential Random Graph Model (TERGM). The TERGM is an extension of the cross-sectional Exponential Random Graph Model (ERGM) to repeatedly observed networks (= panel network data). The following topics are covered: 1) statistical and theoretical foundations of the TERGM, 2) case study examples on the use of TERGMs, 3) conceptual similarities and differences between the TERGM and the Stochastic Actor-Oriented Model (SAOM), 4) preparing data for a TERGM analysis in R, 5) specifying and estimating TERGMs using the xergm package in R, 6) specifying different kinds of temporal dependencies, 7) assessing and comparing the goodness of fit of (longitudinal) network models, 8) out-of-sample prediction, and 9) micro-level interpretation of (T)ERGMs. Basic familiarity with cross-sectional ERGMs using the R packages statnet and ergm is required for participation (as covered by the ERGM workshop in the morning session). Participants should bring their laptops with R and xergm installed.

June 17, 5:30–7:30pm

**NEW** A User's Guide to FEC Data - Navid Dianati, Derek Ruths, and David Lazer

In this workshop, we will introduce a newly disambiguated version of the FEC's campaign contributions database and the open source software designed to produce it. First, we will describe the original database released by the FEC and the challenges in using the raw data for research. Then we will discuss the methodology (and the implemented algorithm) used to disambiguate the donor identities within the database. This will be followed by an overview of the validation methodology used to evaluate the precision and recall of this process, and results are presented. Finally, we introduce a Python API that allows researchers easy online and offline access to the database via custom queries, and walk interested participants through a number of hands-on examples and exercises. In the end, those interested in contributing to the further development of the software will get a more detailed tour of the algorithm and the source code.

June 18, 9:00–12:30pm

Network visualization with R - Katherine Ognyanova

This workshop will cover basic and advanced network visualization techniques using the R language for statistical computing. Participants should have some prior knowledge of R and network concepts. The course will provide a brief overview of network formats, focusing on their structure and representation in key R packages. The training will include a step-by-step guide describing (through series of examples) the path from raw data to graph visualization in Statnet, igraph, and qgraph. The workshop will also provide tips on exporting network data from R to other visualization platforms (e.g. Gephi, NodeXL). The advanced portion of the workshop will introduce dynamic/animated network visualization using the ndtv package. This section will also touch on ways of converting simple graphs in R to interactive JavaScript/d3-based visualizations for the Web.

Latent Space Modeling for Social Networks - Jason Morgan

This course will provide a theoretical and applied introduction to latent space models for social networks. In addition to providing an introduction to the specification and estimation of the standard cross-sectional latent space model, the course will also provide detailed discussions of model identification, diagnostics, and interpretation. More advanced models that provide for node clustering and sociality effects will also be introduced. Finally, a brief discussion of newly developed methods for analyzing longitudinal models will be provided. It is strongly recommended that workshop participants have a working familiarity with R, which will be used extensively throughout the workshop.

June 18, 1:30–5:00pm

Community Detection - Scott Pauls

Community detection in networks is a basic tool for both simplifying and de-noising data. For several standard methods for detecting communities, we will explore the background assumptions, applicability of the methods, and computational considerations. To keep the discussion from being overly theoretical, we will apply methods to running examples using empirical political science data.

Generalized Exponential Random Graph Models: Inference for Weighted Networks - James Wilson

The exponential random graph model (ERGM) is a popular and flexible tool for statistical inference with network data. However, a major limitation of the common formulation of the ERGM is that it can only be applied to networks with dichotomous edges. In this workshop, we will present a flexible distribution on weighted networks that generalizes the ERGM to networks with weighted edges. Through the use of Markov Chain Monte Carlo methodology, we show how to efficiently simulate and estimate these generalized exponential random graph models (GERGMs). Through a numerical study with both simulations and real applications, we highlight the capability of the GERGM, and show how one can use this framework to avoid degeneracy issues that commonly arise in the use of standard exponential random graphs. All applications will be illustrated using built in functions from the R package xergm.