Introduction to Network Analysis

The project I chose to explore is the Introduction to Network Analysis by Thomas Padilla and Brandon Locke. This post is mainly focused on guiding how to start analyzing and creating a network analysis using Gephi, a data visualization tool.

When the data are imported, Gephi shows a default network visualization. This guide provides an example source of network node and edge data, so Gephi will render the network based on the provided nodes and edges.

The visualization above follows the provided data, but we can add adjust and emphasize the relationship of the nodes. For example, there are three categories of the nodes in the network nodes: Individual, session, and location. By distinguishing the categories, we can adjust the relationships of the edges, connecting each of the related nodes such as individuals and location, or individuals and session.

After the adjustment of the network, we can check that the types of nodes are distinguished with different colors. Also, it shows that the individual nodes are linked with the location node and session node, and the session node is at the center of the network visualization from the image and the individual nodes surround. Thus, we can infer that the individuals from different locations (or universities) participated in the linked session node.

Besides the adjustment of categories, we can also apply the degree metrics to see the significance of each nodes. Here, Padilla and Locke explain that choosing ‘Out-Degree’ would emphasize the sizes of the location nodes and session nodes by the number of the linked individual nodes to them, and we can predict the relative sizes of the sessions and maybe the locations (universities).

Reading over Padilla and Locke’s guide, I had an impression that Gephi is a simple and powerful tool to visualize the network of data. It also supports useful features such as distinguishing category and degree metrics. However, besides the usefulness of the software, the post was basically introducing Gephi and was using an example data, so I had to make a conjecture about the relationship of the categories of the nodes, which was the hardest part to think of.