JudaicaLink, a Knowledge Graph for European Jewry

JudaicaLink is a knowledge graph to search for entities (such as persons, places, dates and other) of european jewry. The system is open source and open access. Currently the knowledge graph includes 9 datasets, like from the National Library of Israel, the German National Library, YIVO and other. It consists of about 58.111 entities.

The knowledge graph is based out of the Hebraica- and Judaica collection of the University Library Frankfurt am Main. It is the largest in Germany and one of the biggest worldwide. It hosts the collection Compact Memory, which consists out of 456 Jewish Journals and Newspapers from 1768-1938. These periodicals represent the complete religious, political, social, literary and academic spectrum of the Jewish community and the "Science of Judaism", thus constituting a major source for the research on Judaism in the Modern Age. The journals have been digitized and gone through a process of machine learning algorithms to extract, match and link entities and contribute them to the knowledge graph. From Compact Memory about 1.9 billion entities could be extracted and matched to JudaicaLink. From these entities entity pages have be generated to publish the information on the web.

This lecture on Digital Humanities uses JudaicaLink as example. It shows how the whole process is being done, from digitization to entity finding and entity matching through machine learning algorithms. One of the challenges for example, is to find entities in different languages (German, Hebrew, English, Russian, etc.). We show how scholars can use JudaicaLink to find information and search entities. But also the technical process on how the data is being generated.

JudaicaLink developed by the Specialised Information Service (SIS) Jewish Studies at the University Library at Goethe University Frankfurt am Main and the Hochschule der Medien, Stuttgart. It is funded by the German Research Foundation.