Session 10
Data analysis: geovisualisation
Hodel Taught by Tobias Hodel.
In this session
- Space and place in humanities data
- Hands-on with the DARIAH Geo-Browser and Palladio
Mapping humanities data
- The gap between a place and a coordinate, and everything that falls into it
- Geocoding and gazetteers: GeoNames, Wikidata, Pleiades
- Historical geography: borders move, names change, places disappear
- The base map is an argument. A modern map under early modern data is a claim.
- Time plus space: how to show change without implying a smooth process
Case study: the geography of tragedy
We work through (Craig 2025) together, because it is a rare example of a geospatial humanities project that publishes every stage of its workflow — and so lets us see the decisions that a finished map normally hides.
The chain, end to end
| Stage | Where it lives |
|---|---|
| Source texts, TEI-encoded, with the printed editions they derive from | Zenodo dataset |
| Place allusions extracted, georeferenced, mapped | TLCMap layer 2246 |
| The argument | DSH article, open access |
This is the course in miniature: encoding → modelling → analysis → publication. Open all three side by side.
What to look at
- In the TEI: how is a place allusion marked up, and what does the encoding decide that the map then inherits? An allusion to “Rome” in a play is not a coordinate until somebody makes it one.
- In TLCMap: the workbench holds the data and the visualisation. Switch between the map and the underlying table. Which points would you challenge?
- In the article: how does Craig move from a distribution of dots to a claim about Shakespeare? Where is the inference, and is it carried by the map or by the reading?
The awkward questions
- Every place in the sample is a place named in dialogue. What geography does that method make invisible?
- The plays are early modern; the base map is not. What does that mismatch do to the argument?
- Eleven Shakespeare plays against eleven randomly selected others: is that a control group?
Hands-on
Working from your Wikidata CSV (Session 5) or from any other place-bearing dataset:
- DARIAH Geo-Browser — georeferencing a table, and animating it over time
- Palladio — map, graph, gallery and table views over the same data; the point is the switching between them
- Deliberately break something: map an ambiguous place name and see where it lands
Reflection
- Which of your points are wrong, and how would a reader ever know?
- What does the visualisation make it impossible to say?
TipIn parallel in the Lab
The DH Lab does data cleaning and reconciliation with OpenRefine — the tool that fixes exactly the place-name problems this session exposes.
Reading for Session 11
An introduction to machine learning (to be announced).
References
Craig, Hugh. 2025. ‘The Geography of Tragedy: Mapping Allusions to Places in a Set of Shakespearean Plays’. Digital Scholarship in the Humanities 40 (3): 747–61. https://doi.org/10.1093/llc/fqaf055.