See below for the preliminary schedule of workshops to be held prior to the conference on June 5th and 6th. The workshops are organized into two suggested tracks, but you are welcome to mix and match as you see fit.
(suggested for beginners)
Morning:
Bomi Lee - Introduction to Network Analysis in R
Afternoon:
Philip Leifeld - Introduction to Discourse Network Analysis
Morning:
Sarah Shugars - Network modeling
Afternoon:
Katya Ognyanova - Network visualization with R
(suggested for more advanced participants)
Morning:
Philip Leifeld - Introduction to Relational Event Models
Afternoon:
Shahryar Minhas - Introduction to Latent Factor Models for Networks
Morning:
Skyler Cranmer - Introduction to ERGM
Afternoon:
Skyler Cranmer - Introduction to TERGM
This workshop will cover a comprehensive introduction to network analysis using the R programming language. Participants will be introduced to basic network concepts, network data management, and network statistics (e.g., centrality, triad) using the igraph and Statnet packages. Following some introduction slides, participants will have chances to practice using simple toy datasets and real-world political science data (e.g., alliances). The last part of the workshop will cover network configurations in signed networks (with both positive and negative ties) using the signet package. I assume that participants have no prior knowledge of network analysis and only basic familiarity with R.
Political networks often develop over time with instantaneous ties, often measured with time or date stamps. For example, international conflict, discourse networks in public policy, legislative co-sponsorship ties, email communication, treaty ratification, and parliamentary vetoes are all examples of network ties that can, in principle, be measured on a fine-grained time scale. Relational event models are a class of likelihood-based parametric models with origins in survival analysis, discrete choice analysis, and network analysis. They can model sequences of network ties as a function of exogenous data and functions of the past observed network sequence. This introductory workshop provides a conceptual and practical overview of relational event models and an overview of software solutions in R with a view to possible applications in political science.
In political science and public policy, we often study how political actors relate to concepts in text data, such as news media or Congressional testimony. For example, actors can use frames, exchange arguments in policy debates, express policy beliefs, or contribute narrative elements to discourses. This can naturally be modelled as a temporal network. Discourse network analysis is a toolbox for analyzing this bipartite, temporal network data structure. It proposes several kinds of data transformation to track changing actor coalitions and clusters of concepts in political processes, methods for analyzing polarization in temporally changing discourse networks, methods for distinguishing central from redundant concepts, ideological scaling of actors and concepts, temporal embeddings and phase detection in temporal discourse networks, and statistical modeling of discourse networks as relational event sequences. A software package with a GUI, Discourse Network Analyzer 3.0, and its companion R package, rDNA, are available to support these analyses. This introductory workshop will introduce the software and provide an overview of the different techniques suitable for the analysis of discourse networks.
This workshop will offer an introduction to latent variable models, with a special emphasis on latent factor models (LFMs). We will explore how models in this category serve two main purposes: performing regression analysis on dyadic data and, from a measurement standpoint, deriving lower-dimensional representations that illustrate the relationships between actors. Our discussion will include practical examples from political science to illustrate these objectives. Finally, we will examine how effectively these model implementations generalize to longitudinal networks and highlight important considerations for using them in inferential tasks.
Scholars face numerous conceptual and methodological questions when going from research question to network analysis. For example, decisions made about data collection will influence how a network’s nodes and edges are defined. Once a network is constructed, researchers face even more questions in determining what approaches are most appropriate for analyzing and interpreting that network. This workshop will focus on breaking down decision points at every stage of the network modeling process. Through real-world examples drawn from across subfields, we’ll discuss how these choices are made and justified. Geared towards beginners with no programming experience, this workshop will include hands-on examples in the R programming language in order to demonstrate how these decisions might be implemented as part of a research pipeline.
This workshop introduces participants to a powerful and popular technique for inferential network analysis: the exponential random graph model (ERGM). This workshop focuses on the statistical methodology underlying the ERGM, the model’s implementation in R, and the interpretation of model results based on detailed discussions of real-data examples. A basic understanding of the anatomy of networks and their description (such as the material covered in the “Foundations Track” workshops) is assumed and familiarity with maximum likelihood estimation will be helpful. Participants unfamiliar with R will likely not understand the code portions of the workshop. This morning workshop is the first part of the “Inferential Track” of workshops and the afternoon “Temporal Exponential Random Graph Model (TERGM)” workshop will pick up where this leaves off.
This workshop will cover network visualization using the R language for statistical computing (cran.r-project.org) and RStudio (rstudio.com). Participants should have some prior knowledge of R and network concepts. The workshop will provide a step-by-step guide describing through series of examples the path from raw data to graph visualization in the igraph and Statnet frameworks. The advanced portion of the workshop will touch on dynamic visualization for longitudinal networks and combining networks with geographic maps. We will also discuss ways of converting graphs in R to interactive JavaScript/3D-based visualizations for the Web.
Update (2024-06-05, 6PM): due to circumstances beyond our control, Katya must conduct this workshop online. Participants who registered for this workshop have several options:
Participate in the workshop via Zoom from the FSU campus. Please meet in front of room 208 by 12:50PM and we will walk you to the room with the Zoom equipment.
Participate in the workshop via Zoom from a location of your choosing. If you wish to join using this method, please email us at polnet2024@googlegroups.com for the Zoom link.
Take Skyler's TERGM workshop instead (not recommended unless you are already familiar with ERGMs).
This workshop picks up where the “Introduction to the ERGM” workshop leaves off and introduces the participant to the temporal exponential random graph model (TERGM), an extension of the ERGM for use with longitudinally observed networks. The workshop will introduce the TERGM theoretically, discuss its estimation, show its implementation in R, and discuss/interpret results of a real-data example. Familiarity with the anatomy of networks and their description as well as the ERGM (at least at the level of the “Intro to ERGM” workshop) as well is assumed. Familiarity with maximum likelihood estimation will be helpful. Participants unfamiliar with R will likely not understand the code portions of the workshop.