Learn about how Notre Dame researchers utilized three data repositories to create digital editions for Latin texts.

Any computational study, no matter the tools used, relies on high-quality data to produce high-quality results. Yet many texts are not afforded this quality: they remain undigitized, or their digital versions suffer from a lack of standardization and quality issues. 

This workshop explores the process of creating quality data through digital editions. It uses the set of 32 editions in the Notre Dame Digitized Latin Collection (ND-DLC), created with the support of the Navari Family Center for Digital Scholarship, as its primary example. Throughout this workshop, we will describe the general stages of a digitization pipeline, discuss specific tools and data formats used to expedite the process, and identify current bottlenecks in the digitization process. 

This session will present a complete example of what a textual digitization project looks like, extending from raw PDFs to refined TEI-XML. Additionally, it will detail the work that resulted in a published data paper.

Presenter:

Stephen Bothwell is a sixth-year graduate student in the Natural Language Processing Group. Advised by Dr. David Chiang, Stephen works on projects pertaining to language at various granularities. One line of work concerns the derivation of sound laws in historical linguistics, striving to add interpretability to derived systems of ancestor and descendant language forms through the determination of rules. Another line of work explores stylistics, attempting to computationally detect and analyze rhetorical devices such as rhythm and parallelism.