GLOBALISE data sprint
The logic of lists and tables: creating ground truth for VOC records with complex layout
The VOC archives contain numerous records featuring lists and tables. These range from records typical of commercial activities, such as cargo lists and personnel lists, to other types of information such as eisen (demands), schenkagie (lists of gifts), population census, or even rare examples such as menus. In this datasprint, we will share about different types of lists and tables featured in the VOC archives and explore their potential for new research avenues.
This datasprint focusses on practical digital skills and data preparation methods. We will introduce participants to the basics of working with the Transkribus platform and invite you to join us for a ‘live’ transcribe-a-thon. Lists and tables remain challenging for layout and handwritten text recognition (HTR) models to process, and human inputs are still necessary to improve the transcription model.
Together, we will learn how to extract information from lists and tables to generate new ground truth data, which will in turn further improve the models for the transcription of lists and tables. Participants of this datasprint will gain a first-hand look into the latest advancements in layout recognition technology, gain access to our curated selection of lists and tables, and learn how to apply layout and HTR models to their own data.
Programme
13:30 – 13:45: Welcome & Introductions
13:45 – 14:30: Datasprint part 1
14:30 – 14:40: Break
14:40 – 15:00: Presentations, including Q&A
15:00 – 15:45: Datasprint part 2
15:45 – 16:00: Closing discussions
This is a hybrid event. Please indicate if you will attend online or in-person in Amsterdam. Further details, including link for online attendance, will be provided closer to the event date.
