eng: This work explores professional and crowd-sourced datasets about person-to-person networks in the context of art history. Motivated by historical examples to represent developments in the arts in diagrammatic form, it seeks to explore 1) if and how data-driven network visualizations can support the contextualization of art- works in virtual presentations, 2) if and which large-scale art history narratives are embedded in extensive network data and 3) if and how they differ across multiple datasets.Social relationships encoded between person records of an institutional dataset, the Getty ULAN, are integrated with artwork metadata in order to create a virtual 3D art gallery environment based on an automatic network layout. Feedback from domain experts shows that it provides an interesting and novel way to explore art history in the digital realm. An overall analysis of the ULAN network reveals a contiguous chronological structure spanning multiple centuries, whose multitude of historical interactions yields a large- scale narrative that correlates with related scholarly views to a certain extent. The comparison with similar content derived from Wikipedia biographies reveals significant overlap as well as structural commonalities and differences between the ULAN network and corresponding hyperlink networks in various Wikipedia language versions, whose separate comparison reveals cultural self-focus bias regarding the coverage of art history biographies but also basic agreement on the fundamental developments in Western art. Wikidata is introduced as an alternative means to identify biographies and other Wikipedia articles relevant to the domain of art history, demonstrated by the analysis and visualization of a bi-partite network of articles about persons and art/architecture styles. Clustering co-occurring occupations in Wikidata persons records is eventually introduced as additional approach to identify domain-specific biographies, demonstrated by the example of an additional artist network and the visualization of networks from other domains.The main contributions of this work are as follows. It identifies, analyzes and compares contiguous and chronological bottom-up structures in large-scale biographical networks extracted from both professional as well as crowd-sourced datasets related to art history, showing that they represent major developments across the ages and complement each other in a way that their combination yields a more global view on the history of art. Moreover, it demonstrates the benefit of integrating different but related cultural heritage datasets for analytical purposes such as their mutual quantitative comparison and the identification of gaps in the data. Last but not least, it suggests new ways to identify domain-specific groups of persons in general-purpose collections of person records.
|