Yale University Art Gallery
Enhancing the Visitor Experience of an Online Collection
The mission of the Yale University Art Gallery is to encourage appreciation and understanding of art and its role in society through direct engagement with original works of art. The gallery’s online collection contains more than 160,000 works, spanning more than 4,000 years of human history. Many of the images in the collection are available for use as part of the University’s Open Access Policy for public domain works.
Goals for Improvement
Having launched their new website in Drupal 7, the Gallery was looking to improve its functionality by making the site easier to use and by simplifying their nightly collection synchronization.
The primary high-level goals for the project were to:
- Streamline collection imports. Provide reliable synchronization from TMS source data to Drupal content nodes
- Present a searchable collection. Provide end-users with relevant search results across content types and within collections based content
- Implement faceted search. Provide end-users the ability to refine search results based on key facets which describe the search data
- Include sortable results. Provide end-users the ability to sort search results when searching collection objects
- Implement predictive autocomplete for collections search. Provide end-users the ability to see a set of auto-complete options of artist names when searching collection objects
- Update the event calendar. Provide end-users with the ability to use the site calendar and event system to find relevant upcoming events
Like many museums, the Gallery strives to put its extensive collection online. At the time of this writing, the Gallery had only 3,800 works on view to the public, with another 160,000 either in storage or available only by appointment. With the collection online, visitors, artists, and researchers can take a deeper dive into the entire catalogue.
For its collection management, the Gallery uses The Museum System (TMS), a Windows desktop application used by more than 800 institutions worldwide. TMS is a proprietary software system that provides its own bespoke web publishing system, and the Gallery wanted a greater level of presentation control than afforded by that system. In particular, the categories used internally at the Gallery needed to be altered when used on the public website.
When Palantir first engaged with the Gallery in September 2014, they were exporting the entire TMS collection nightly and importing it into the Drupal site using the Migrate module. The process was lengthy and error-prone, so our first task was to make the data import more scalable by having the nightly import only respond to records changed in the TMS system since the last import.
The Migrate module supports this feature with the concept of highwater marks. Highwater marks map the imported data to a sortable key in the source data -- typically the last updated date -- and instructs Drupal to only return content that has been changed since the last import. Doing so cut the speed and increased the accuracy of the nightly import, which formed the basis for all further site improvements.
To search the collection, the Gallery installed an Apache SOLR instance, creating search repositories for site content and for the collected works. This dual approach allowed the creation of two distinct search experiences. Overall site search returns collection objects, events, and other content, while the collection search returns a faceted list of collection objects.
Using the Search API module as a base enhanced by the Facet API module, we can provide a rich search experience similar to what one might use on Amazon.com, where search results can be further narrowed by selecting checkboxes that limit the search range. For the Gallery, a search can be filtered by six distinct filters: Department, Classification, Date Made, Culture, Image Available, and Availability in the gallery.
Using the Facet API Slider module, we constructed a custom date widget that allows the user to select a start and end date using a doubled-handled slider.
To create predictive search results, we created a custom autocomplete callback to handle the requirement that searches should default to pulling back artist records. (Technical note: since the search field is targeted to the SOLR full text index, the Search API Autocomplete module turned out not to work for this use-case.)
The Gallery hosts numerous public exhibitions and events throughout the year. They play host to a number of traveling exhibitions and need to keep an overall calendar that includes events syndicated from Yale’s central calendar service.
Like the collection search, the calendar includes facets to help visitors narrow the list of results. The site also tracks upcoming and past exhibitions, serving as a showcase of the breadth of the collection.
Like many organizations, the Gallery has complex publishing needs and a small staff. A major outcome of our work has been to streamline the production of the site so that it can be automated based on work already being done in external systems. Collections, exhibits, and artist information all flow through the TMS migration, allowing curators to do their work once and not spend extra time managing the website directly.
Streamlined updates aside, the primary result has been the improved experience for Gallery patrons. By improving the search and calendar functionality, visitors are now able to engage easier with the original artwork provided by the Gallery, both online and in-person at events.