Frank Schweitzer - Cultural Data Analytics - The Bigger Picture: Documents -> Social Networks -> Knowledge Graphs

When: 2022-05-09 10:00-12:00 (Tallinn time)
Where: CUDAN Open Lab room A108 and online via zoom.

The event is public via zoom:

Abstract – As Herbert Simon states it, “a wealth of information creates a poverty of attention”. Often, this wealth of information is not even realized. Thousands of hand written documents have been collected in state archives over centuries, but only a few specialists are able to decipher them. Millions of printed matters are stored in libraries, but noone reads or borrows them again. For instance, the correspondence of Heinrich Bullinger, a renowned Zurich reformator from the 16th century, contains more than 10,000 letters from and 2,000 letters to Bullinger. Only about half of them are edited so far.
The Swiss parliament possesses the most complete record of all its debates, bills and issues for more than 127 years, but there is no systematic overview of how they are linked together. Transcribing handwritten (see the READ-COOP) or OCRing printed documents does not really improve the situation. It mostly adds to the “known unknown”, the pile of those documents that exist but are never utilized because we have developed neither research questions nor tools to analyze them.

The network approach builds on the belief that knowledge can be extracted from linking documents by means of their meta information rather than spelling out their content. In my talk, I will illustrate the potential of this methodology using the two mentioned examples of the letter correspondence between 16th century reformators and the records of the Swiss parliament. How can we extract social networks, measure the importance of individuals and detect their friendly or adversarial relations? How can we distinguish meaningful from random interactions and consider preferences for interactions in network models? How can we build large-scale knowledge graphs? The talk will address these questions with a focus on methodology. It spares the details, to catch a glimpse of how the wealth of isolated information can be turned into a new attention for relationships.

Frank Schweitzer Chair of Systems Design, ETH Zürich, Switzerland

European Commission logo Tallinn University logo