Dataset:

The database of manuscript items contains 1306 records, 1187 textual units, 676 records as Persons, and 274 Imprints. It can be viewed as a basic list of records, and each item can also be explored or searched individually, or visualised geographically as a set of relations through selected filter options.

The dataset is largely based on the online catalogue of the International Collection of Digitised Hebrew Manuscripts (http://web.nli.org.il/sites/NLIS/en/ManuScript/), whose records have been expanded, updated, translated and/or transliterated, linked to existing online resources and data repositories when possible, and hierarchically organised. 

For manuscript items, the relevant sources were selected on the basis of palaeographical identification (all texts, with some individual exceptions, are written in the Ashkenazi script), place of production (if recorded in the object), and subject matter recorded in the NLI catalogue (that is, with the subject predefined as kabbalah, or with the keyword kabbalah contained anywhere in the item description).

Database Structure:

The dataset has been ingested into nodegoat—an online software research environment. The environment allowed us to model data flexibly, import data, and export datasets in CSV/ODT format. It also delivered geographical, chronological, and social visualizations of our stored data. 

In nodegoat, our data was structured into objects (physical items) and their subordinate sub-objects (textual contents). The sub-objects contain relationships to other objects but also store various properties related to their dating, geographical location, or persons involved in their history. The vocabularies used for specific elements within our database partially follow the nodegoat model but were adjusted to fit our needs.bWe have incorporated the geonames ontology for place names (http://www.geonames.org/ontology/documentation.html). Personal names have been matched to the Virtual International Authority File (VIAF, https://viaf.org/).  

Please get in touch with our team if you would like to suggest corrections. You can access the full dataset and cite all versions by using the DOI 10.5281/zenodo.15033262.