Faceting

From UA Libraries Digital Services Planning and Documentation
Jump to: navigation, search

Our researcher study on "What Do Researchers Need?" showed us that in order to locate content in a large database, researchers need faceting, limiting, and browsing capabilities.

In order to develop this service, we need an underlying strata of metadata that will serve. For example, if you want patrons to be able to drill down and select only the results that fall within a particular decade, then the metadata must already be indexed that way. Likewise, if you want patrons to be able to drill down and select only the results that fall within a particular topical grouping, then the metadata needs to be there to support it.

Whatever results you want from the finished product are often dependent upon the metadata.

We already have tons of content in Acumen, so we analyzed the existing metadata to find out what kinds of groupings we could make that would be useful.

One of our guiding considerations is that whatever dropdown menus we provide should have no more than 15 (preferably 10 or fewer) choices, from a usability standpoint.

So, the metadata librarians are grouping the genres into categories and subcategories... and then will do the same for topics. The idea is that for each facet selected, the most general groupings are seen first, and the user drills down to more and more specific groupings, to limit his/her search results.

We envision the following facets (click the links for more information on how we're doing this):

and we may add more after that. The current schema for facets shows the details of the XML we add into MODS extension. (This file is found on the server in /srv/www/htdocs/digital/schemas/.) Whenever the facets are updated (or first created) the updated_f timestamp is updated (or created).

The first time facets are introduced into a MODS extension, the namespace for the facet schema is also added. Also, the recordInfo:recordChangeDate is added and/or modified when the record is changed. Thus, by comparing the most recent recordChangeDate and the updated_f value, the script can identify whether the file has been changed since the facets were last updated. If it has, then the faceting script re-analyzes the file and regenerates the facets.


The geographic facets, however, are supported by making use of MODS hierarchicalGeographic fields.


Our web programmer is handling the SOLR indexing and interface part (thanks, Will! You're the best!) The links above provide more information about how the encoding takes place in the MODS.