New Visualizations of Data: Coincidences and Regressions through Network Analysis

Graphs are widely used to represent social structures and visualize correlations between variables. While traditional methods like trajectory analysis and structural equation modeling focused on quantitative data, recent tools such as the R package netCoin and the Stata ado coin now enable the visualization of connections between qualitative variables. These tools combine multivariate techniques like linear and logistic regression with network analysis, offering new ways to explore qualitative data.

Network Coincidence Analysis (NCA) provides a novel approach to analyzing survey data, particularly for examining qualitative variables. In this method, nodes represent categories of selected variables, and links illustrate relationships between them. This technique is especially valuable for identifying response profiles based on socio-demographic variables. Beyond correlation measures, NCA can estimate log-linear models to study multivariate relationships, including interactions, thus offering a deeper understanding of complex data patterns.

One of the strengths of this approach lies in its interactivity. The tools allow users to filter and manipulate the data visually, selecting items based on size, attributes, or centrality. This interactive element enhances both the user experience and the presentation of results, making it easier to engage with and interpret complex relationships between variables.

This presentation will first explain the statistical foundations behind NCA and then demonstrate its application using data from international surveys such as the European Social Survey. These case studies will show how visual tools like netCoin can transform the exploration of qualitative data, offering an innovative and experiential approach to research, in line with the session’s focus on creative formats for presenting visual research.