GIGI & The Bunraku Collection

For this blog post I looked at two different repositories: The Graphic Interface of Gathered Images(GIGI), which is a special collection of the American Antiquarian Society, and the Bunraku Collection, which is part of the Columbia University archive. Both of these collections are comprised of digital images and are freely accessible. In this post I will describe each of the collections individually and then present my conclusions.


First, I will discuss the GIGI collection(http://gigi.mwa.org/). This is a special collection of the American Antiquarian Society that is not integrated with their online catalogue. Upon entering the site the user is immediately presented with a display of images on the right and a sidebar on the left. The images displayed do not appear to be of any significance, but I suppose they are intended to give the user a sampling of what is stored in the collection. Within the sidebar on the left is a keyword search box, a browse tab with subject terms, and an advanced search tab. For the purposes of this post I will focus on the browse tab.

Within the browse tab there are subject headings such as African Americans, Cartoons, Maps, Photography, United States History...etc. Clicking on one of these terms will display all the items in the collection that are associated with that term. For example, clicking on “manuscripts” will display the 47,861 images that have been assigned that label. Scrolling through the first page of results for manuscripts it seems that there is quite a bit of metadata linked to each item. However, if one looks over the first page of results for photographs the lack of description at the item level is obvious. This marked difference in the amount of descriptive data is indicative of the institution's priorities and perhaps exemplifies the negligence of photographic materials in the archival context.

Browsing through the results of the first page I was not able to find an image that had any substantial metadata associated with it beyond the keyword term “photography” or in some cases the format or process (e.g. Daguerrotype, stereograph...etc). Fortunately, as I continued browsing through the photography collection I was able to find items with more than just keyword tags. The standard fields in use appear to be: Title, Author-Creator, Date, Call Number, BIB ID, Keyword, Notes, Related Names, and Filename. The fact that not all of the images in the collection contain this basic metadata is problematic in terms of accessibility. However, based on the images that had this information I did not see any glaring problems. The majority of the keyword tags associated with each image are copied directly from the record. This seems to be a decent way to create tags, especially if it is done automatically. The only problem is that records with little to no metadata result in inadequately described images that are difficult for users to discover.


The Bunraku Collection, is a photographic collection housed at the C.V. Starr East Asian Library of Columbia University. The digital version of this collection is accessible online through the Columbia University Library website(http://www.columbia.edu/cu/lweb/digital/collections/eastasian/bunraku/), but it is essentially a special archival collection. From the home page of the site the user can search using keywords or browse by collection or image type. There is also a very detailed finding aid that can be viewed by mousing over the “About the Collection” link at the top of the page.

The first thing I noticed about this collection is that it can be browsed in two different ways. Browsing by image type, as in the GIGI database, presents the user with a highly unorganized mass of images that are not easy to sort through. Users can click on individual records and be brought to more specific and organized amalgamations of images via keyword tags, but this is truly browsing and not really what a user with specific queries wants to be dealing with. Therefore, by allowing users to also browse by collection I think the bunraku collection is superior to GIGI.

By browsing by collection users can narrow down the image results to suit their needs. For instance, if I only want images from a specific play I can click on “Play Title”, find the specific play(titles in English and Japanese) and then instantly narrow down the image results. It is also possible to do this for specific performers, characters, roles...etc.

The success of this browse functionality is heavily dependent on the description at the item level and I think the archivists here have done an adequate job of supplying each item with proper access terms. Each item is described using the fields: Title, Production, Scene, Performers, Characters, Kashira, Subjects. When applicable descriptions are provided in English, Romanized Japanese, and Japanese. The use of these fields is what allows for users to browse the bunraku collection in a more meaningful way than the GIGI. Although, the one area in which I think the GIGI excels is the notes section. What I noticed in the GIGI records is that the notes were much more detailed and this lead to more descriptive keywords terms. However, I feel like adding notes to the bunraku images would, in most cases, be superfluous given the precision of their other assigned terms.


Overall, after looking at these two repositories I feel that the bunraku collection does a much better job of describing and linking its contents in a meaningful way. The descriptive terms used in the bunraku collection are specific to the collection and indexed in a way that facilitates access for the user. It would be unfair to say that the descriptive terms used in the GIGI collection are inappropriate, but it is clear they do not provide that same level of precision and thus the collection as a whole is not as accessible.