Network Analysis With Gephi

Network visualization tools, algorithms, and custom-built interfaces can help visualize and analyze relationships in datasets that might otherwise be invisible, as exemplified in the Virtual Texts, Tudor Networks, and Linked Jazz projects (Arnhert & Arnhert, 2023; Cordell, 2015; Lange & Pattuelli, 2014). In the Virtual Texts project, for example, researchers used network graph tools to visualize text sharing within a dataset of 19th-century newspapers and magazines, showing how specific texts were copied and recirculated across newspapers and magazines within regions and over time.

Interactivity with these projects can also reveal network relationships between nodes and edges at individual, network, and temporal levels. The Linked Jazz project offers users robust interactivity at the individual musician level and for visualizing relationships between jazz artists and other artists (Lange & Pattuelli). The Tudor Networks project visualization allows users to see how state letters traveled across time and space and to visualize the movement of a person or institution’s communications as bursts, showing their reach. We can also see the scale of communications by the size of the symbolic circle used (Arnhert & Arnhert, 2023).

Gephi, a network analysis tool, exemplifies how network tools can extend a user’s interactions with a project and enable researchers to understand relationships between edges and nodes in datasets in new ways. Using the Gephi tool with the “Early African American Film Database, 1909-1930 dataset revealed additional capabilities with network graph tools, including the ability to visualize networks differently. Tools like the Fruchterman-Reingold algorithm can untangle network graphs and enhance their visibility by dispersing groups and creating space around larger nodes, making more densely connected parts of the network more visible. Also, tools like the modularity report highlight internal subdivisions within a network into distinguishable groups. Gephi tools, such as the Multimodal Networks Projection tool, also allow researchers to convert bimodal networks into unimodal networks, altering how links and edges connect for even greater visualization.

Leave a Reply

Your email address will not be published. Required fields are marked *