Home - dkglab/fall-of-rome GitHub Wiki

Graphing the Material Fall of the Roman Empire

Archaeological datasets have revolutionized the study of the Roman Empire’s fall, illuminating its everyday realities.

But data-driven historical research faces challenges:

  • data collection is labor-intensive,
  • data is small but complex, and
  • historical arguments emphasize narrative over database.

By publishing historical data and argumentation as a knowledge graph, we are not trying to “solve” these challenges but to work with data in a way that acknowledges them as constraints.

The data

For his dissertation Dr. Gruber compiled a MetaSource from archaeological data to build an argument about Rome’s fall in Spain and Portugal.

Our project is converting this data into a knowledge graph, a network of relations among:

Read about the data processing pipeline.

Short-term goals

  • Express Dr. Gruber’s work as a knowledge graph
  • Designing a workflow to transform his relational tables into graph data
  • Interlink this graph with other relevant scholarly knowledge graphs
  • Refine the data model by creating a Web-based publication with interactive visualizations and a hypertext presentation of Dr. Gruber’s argument

For more on the latter, see PresentingTheArgument and Architecture.

We translate these goals into stories and tasks by playing the PlanningGame.

See the Tools we plan to use.

Longer-term goals

  • Explore workflows for transforming Dr. Gruber's source materials (such as archaeological field reports in PDF format) directly into graph data

  • Evaluate machine learning techniques for automating data creation and editing while maintaining high confidence

Related work