{"id":122,"date":"2019-10-29T16:08:10","date_gmt":"2019-10-29T20:08:10","guid":{"rendered":"http:\/\/sites.nd.edu\/marble\/?p=122"},"modified":"2019-10-30T16:51:23","modified_gmt":"2019-10-30T20:51:23","slug":"metadata-connections","status":"publish","type":"post","link":"https:\/\/sites.nd.edu\/marble\/metadata-connections\/","title":{"rendered":"Metadata connections"},"content":{"rendered":"

By Hanna Bertoldi <\/p>\n

What is metadata?<\/strong><\/h2>\n

I am constantly thinking about metadata, or the data that describes other data. <\/span><\/p>\n

While most people may not often think about metadata, everyone who uses search tools benefits from clean, well-structured metadata. <\/span><\/p>\n

Think about a music file. Its metadata might contain the artist\u2019s name, song title, album title, and year it was released. Spotify uses this metadata to help you find the songs you want and also to suggest music you might like. Some of these terms may be hidden from you, such as terms that describe the mood or genre of a song, but they all work in the background to create easily browsable categories. <\/span><\/p>\n

\"Screenshot

Figure 1: Spotify browse page<\/p><\/div>\n

Museums and metadata<\/strong><\/h2>\n

Like Spotify, museums also create and use metadata, but, instead of music, the metadata describes pieces of art. <\/span><\/p>\n

As its best self, museum metadata has the potential to bridge gaps between cultural heritage institutions and provide access beyond virtual and physical barriers. This unbounded potential can only be leveraged by using accurate and standardized metadata. <\/span><\/p>\n

But, when I think about museum metadata, I mostly think about how messy it is. (Apparently, <\/span>Spotify thinks about this too<\/span><\/a>). In the screenshot below, the fields \u201cCreation Date\u201d and \u201cCentury\u201d reveal some examples of data inconsistency and redundancy. For instance, you can see that at one time items without an identifiable creation date were marked with \u201cno date\u201d whereas others were marked with the abbreviation \u201cn.d.\u201d In addition, the same date is repeated in both the \u201cCreation Date\u201d and \u201cCentury\u201d fields for many of these objects. These kinds of metadata errors can prevent database users from finding objects that meet their search criteria.<\/span><\/p>\n

\"Screenshot

Figure 2: Screenshot of creation date export from EmbARK, the museum’s collections management system<\/p><\/div>\n

Organizing the \u201cmess\u201d with controlled vocabularies<\/strong><\/h2>\n

Controlled vocabularies are a critical element of creating consistent metadata. A controlled vocabulary is an organized arrangement of words and phrases used to index content and to retrieve content by browsing and searching.<\/span><\/p>\n

If we wanted to build an application where you could search every artwork in the world, every arts organization or collector would have to agree to use the same vocabulary, or a \u201ccontrolled vocabulary,\u201d to describe their objects. We couldn\u2019t have the Getty Museum using \u201cClaude Monet\u201d while the Louvre was using \u201cOscar-Claude Monet\u201d to describe the artist that painted water lilies in the nineteenth century. They would have to agree on the same version of the artist\u2019s name. <\/span><\/p>\n

Beginning in the 1980s, the <\/span>Getty Research Institute<\/span><\/a> started developing controlled vocabularies for cataloging visual arts and cultural heritage.<\/span> The terms and associated information in the <\/span>Getty Vocabularies<\/span><\/a> are valued as authoritative because they are derived from published sources and represent current research and usage in art history and cultural heritage communities.<\/span> <\/span><\/p>\n

In short, the Getty Vocabularies provide a convenient system for mapping data, without spending a lot of wasted time reinventing what already exists. <\/span><\/p>\n

\"Database

Figure 3: Searching EmbARK, the museum’s collections management system, by subject terms<\/p><\/div>\n

Applying the Getty Vocabularies to Snite Museum collections<\/strong><\/h2>\n

Now, back to the messy data part. <\/span><\/p>\n

The Snite Museum of Art<\/span><\/a> has been thinking about subject retrieval for quite a while. Perhaps as early as 2010, Snite curators began cataloging works of art with subject terms or keywords. This metadata allows curators to search for things about an object, such as who or what is depicted. Unfortunately, these terms were used without standardization. <\/span><\/p>\n

Building off of the curators\u2019 previous work, my goal as Collections Database Coordinator was to transform these subject terms or keywords so that they fit within the Getty Vocabularies. On a superficial level, this process has helped to clean up things such as errors in spelling and inconsistencies with formatting and style. On a deeper level, the process created relationships among objects and increased discoverability.<\/span><\/p>\n

\"Photograph

Figure 4: Unidentified photographer, Geisha in a Ricksha, Japan, ca. 1880-1890, albumen silver print with applied color. Snite Museum of Art, University of Notre Dame. Acquired with funds provided by Robert E. (ND \u201963) and Beverly (SMC \u201863) O\u2019Grady, 2008.054.002.<\/p><\/div>\n

Such immediate improvements in discoverability are a result of the way Getty Vocabularies are structured: as hierarchical thesauri. <\/span><\/p>\n

Getty Vocabularies are <\/span>hierarchical <\/b>to allow similar objects to be grouped together. For example, you want to discover all of the works in your collection that depict transportation. <\/span><\/p>\n