Unleashing Museum Collections: Three Case Studies

Greg Turner, Australia

Museums often find it difficult to integrate their collections into their other online properties. Several collections systems vendors offer online components, but these tend to be limited in functionality and customization possibilities.

Our approach to this problem has been to harvest the data from a Museum’s collection, using available APIs and export functions, and store it in the database for the Museum’s site. From there, we can customize publishing workflows, search algorithms and provide views and APIs that allow collection data to be integrated within other pages of a museum’s site, or externally.

This paper explores how this approach has been successfully used by three museums with wildly differing requirements. The tools we used are provided by the open-source GLAMkit framework, which is built on the Django web framework.

Case Study 1: Art Gallery of New South Wales

The AGNSW has a collection of ~30,000 works, stored in a Vernon CMS database. We harvest the data whenever a record changes.

The online collection has functions that let users browse by taxonomy, conduct a free-text search or conduct an advanced search. Results are shown using a customizable faceting engine.

The AGNSW specializes in Asian and Aboriginal art. The search is customized to handle demonyms (a search for ‘Dutch’ will return objects from Holland, for example) and alternative spellings of Aboriginal and Asian names.

The first version suffered from performance issues due to technology choices. The paper describes those issues, and how they were resolved.

Case Study 2: National Film and Sound Archive, Australia

The NFSA has collections comprising ~900k items stored across several heterogeneous databases, each of which are harvested and stored in a central database.

By centralizing the online database of collections, we are more easily able to draw connections between records that did not previously exist. We have also added a facility for visitors to the site to create and share lists of items from any collection, or make a request to access items in the list from the archive.

Despite the improvement to the existing services, the centralized search has highlighted some problems with data consistency across the collections. We explore those issues briefly.

Case Study 3: Museum of Contemporary Art, Australia

The MCA’s collection database is also stored in Vernon CMS. The harvest and storage approach is identical to AGNSW, but we provided an additional editing stage in the web CMS, where material of various kinds can be added by the MCA’s digital team before being published. This process is fairly straightforward in Django, but offers interesting possibilities by creating a ‘collection-plus’ information space.

Whereas Vernon CMS has a per-seat license, and is designed to hold authoritative data, the MCA’s online collection can hold that data plus anything else necessary to support the experience of the public. In time, and with the shifting perceptions of curatorial practice, this added information may take on curatorial importance, in which case it is ready to be pushed back into the canonical collection database.