Session 8
Data analysis: networks
Hodel Taught by Tobias Hodel.
In this session
- Discussion of (Weingart 2011)
- Network concepts for humanists
- Hands-on: Gephi Lite with the CSV from Session 5
Networks, demystified
Weingart’s warning is the frame for the session: network analysis is not a way of looking at your material, it is a way of transforming it, and the transformation is lossy.
- Nodes and edges — and the violence of deciding what counts as which
- Directed and undirected, weighted and unweighted
- Bipartite (two-mode) networks, and why humanities data is so often bipartite
- Degree, centrality, components, communities — what each measures, and what each does not
- The most common error: reading a layout as if it were a map
Hands-on: Gephi Lite
Bring the CSV you produced with Wikidata in Session 5. If it is missing or broken, rebuild it — that is part of the lesson.
- Import the CSV as a node list and an edge list
- Apply a layout, and then apply a different one; compare
- Compute degree and one centrality measure
- Run community detection and ask whether the communities mean anything
- Export the graph and a figure for your project site
An extended worked example of analysing a bibliography as a network: https://github.com/RISE-UNIBAS/networks_gephi
Your Zotero library is already a network: authors, co-authorship, shared keywords, citation. This is one of the most direct routes from the assignment to an actual analysis.
The DH Lab introduces Omeka and compares it with GitHub Pages as a publication route.
Reading for Session 9
Underwood, Ted. 2014. “Theorizing Research Practices We Forgot to Theorize Twenty Years Ago.” Representations 127 (1): 64–72. https://doi.org/10.1525/rep.2014.127.1.64