Visualizing Bibliographic Metadata for Art Historical Research

Visualizing Bibliographic Metadata for Art Historical Research

Over the past two years, Corina has been working with colleagues Tomasz Neugebauer (Digital Projects & Systems Development Librarian at Concordia University) and Felicity Tayler (PhD Candidate, Humanities Doctoral Program at Concordia University) on a data visualization project using the bibliographic metadata from the Artexte collection.

Together they presented some of this work and their findings (and open questions) at the Art History Graduate Students Association Conference (AHGSA) in March 2015. Below is the abstract of the presentation:

This paper responds to the crossover in conference topics between network science and art historical research methods. We ask how the visualization of complex networks can be used to generate art historical questions? Our data set is derived from the bibliographic database created by Artexte, an organization with the mandate to comprehensively collect Canadian exhibition catalogues and related international materials. The nodes of this network include the documents in the Artexte collections, connected to each other through edges representing subject cataloguing (keywords), and contributors, such as: artists, writers, editors, translators, critics, publishers, art organizations, etc. The resulting network of 40,000 e-artexte catalogue records contains over 135,000 nodes and more than 320,000 edges. The emerging research questions for this exploratory study include: (1) does the bibliographic metadata network exhibit the properties of complex networks (properties associated with small-world and scale-free networks) found in previous studies? (2) can the visualization of bibliographic metadata as a network contribute to art historical research? (3) is the measure of betweenness-centrality in the complex network derived from the e-artexte dataset useful to art historical research? We used measures of centrality to determine how important a particular node is to the entire structure. Although we initially thought these measures might reflect canonization processes, they seem to point to something else. Rather than mapping who has the most power in the art world because they are exhibited, published or written about most frequently, it seems to map which publications, writers, artists, curators are most important in holding the whole net together. What meaning does this outcome have for art historical methods?