Skip to main content

Is Granularity the Next Discovery Frontier? Part 1: Supporting Direct Access to Increasingly Granular Chunks of Content

Webinar

About the Webinar

The rise of the Discovery System in the library world has helped to streamline searching for end users by providing them with search functionality that more closely resemble search engines like Google than traditional database searches. But with this streamlined search comes added expectations from users about their ability to drill down into content and retrieve more granular pieces of information—anything from book chapters and individual letters to the editor to specific graphs and images could conceivably be retrieved in a more granular search. What is needed to ensure that discovery systems can retrieve and display information below the publication level? What is the role of the content provider and the library in this scenario? How do libraries help end users find and use this content?

So, is granularity the next discovery frontier? In Part 1: Supporting Direct Access to Increasingly Granular Chunks of Content, this webinar will discuss the implications of granular content for user search interfaces and discovery engines.

For a deeper understanding of metadata and metadata standards, NISO published this open resource, Understanding Metadata.
Feel free to acess here on the NISO website: http://www.niso.org/publications/press/UnderstandingMetadata.pdf

For further reference, MJ Han provided the metadata schema, SPINEhttp://diglib.hab.de/rules/schema/emblem/emblem-1-2.xsd 

Event Sessions

Working with Metadata Challenges to Support Granular Levels of Access and Descriptions

Speaker

Myung-Ja (MJ) Han

Metadata Librarian
University of Illinois at Urbana-Champaign Urbana, Illinois

As the library collects and provides more resources in digital format, new units of granularity have emerged in access and description. Digital surrogates or born-digital resources makes it possible for more granular levels of access, e.g., a book can be accessed at chapter level and page level in addition to a book or journal title as a whole. In order to provide these granular levels of access, metadata should be created in granular levels as well. This presentation discusses how the emerging needs of granularity of access and description makes the metadata and cataloging process a highly collaborative work, and suggests a way to design and create a metadata schema for describing granular levels of resources with examples tested at the University of Illinois at Urbana-Champaign Library.

Granular Discovery: User Experience Challenges and Opportunities

Speaker

Tito Sierra

Director of Product Management
EBSCO Information Services

Web-scale discovery services are transforming how users search the library. By combining print and electronic resources with full-text search indexing and relevance ranking, discovery services hold enormous promise in connecting library users to a diversity of high-quality information resources, quickly and easily. The challenge typically isn’t the availability of quality content in a discovery service, but navigating an abundance of content to identify the most relevant content for the user’s specific information need. High quality metadata is a critical ingredient in this process, as it enables intelligent relevance ranking, support for advanced search strategies, and useful post-search refinement options. However user research and log analysis indicates that most discovery service users bypass advanced search options and post-search faceted search refinement options in favor of basic keyword searching to locate content of interest. Delivering effective discovery and access to more granular slices of content requires not only more metadata, but more effective use of metadata and user context to help users locate the specific content of interest. Supporting increasingly granular discovery, while providing the Google-like ease-of-use that users have come to expect, is a major challenge. This presentation describes the user experience challenges associated with supporting more granular discovery, and concludes with several paths forward for future innovation in this space.

From Unstructured Content to Granular Insights

Speaker

Semantic enrichment is commonly used to identify business objects of interest (People, Companies, Products, terminology, etc…) mentioned in documents and augment their metadata to improve search and navigation. But the same technology can also be used to harvest information on business objects themselves. Documents that we index and search are a treasure trove of such business insights that can be extracted as triples and exploited standalone. In fact, a broad range of intermediary approaches exist to leverage content in increasingly granular ways. In this session we will showcase the principles and concrete examples of information applications built along such approaches. 

Event Q&A

Speakers

Tito Sierra

Director of Product Management
EBSCO Information Services
How do you provide an interface for creating the additional metadata? What kind of people enter it? (metadata specialists, or ordinary people?)

(Daniel Mayer, TEMIS): Our approach is to semi-automate metadata creation and provide the option for authors/editors to add or modify metadata. In some cases our customers enable readers to augment the metadata themselves.

From your log analysis of search terms, can you estimate the percentage of searches which are known vs. unknown searches? And what methods do you use for determining that?

(Tito Sierra, EBSCO): We estimate that known-item title searches account for between 15% and 25% of all searches on EDS. That figure is derived from a search log analysis we completed last fall. Regarding methods, we looked at 4000 random queries extracted from EDS transaction logs and manually coded these queries into four categories: Known-item, Topical/Exploratory, Names, Other/Ambiguous. We found that Public Library instances of EDS had the highest rate of known-item title searching at just under 30% of all searches and School (K-12) Libraries had smallest rate at around 5% of all searches. Academic libraries were under 20%. About 5% of queries (“Other/Ambiguous”) could not be coded based on analysis of the query alone and may include some known-item search intent.

Searching across all library resources is technically possible but still limited by the metadata sharing agreements between content providers and discovery service providers. Do you see that changing or evolving?

(TS): This is an important issue. Discussions are happening between discovery service providers in this area. I am optimistic these issues will be resolved in the future.

Also, why are Discovery Service providers so hesitant about sharing the user's search terms? This can go a long way in improving the user experience and highlighting the search hits for them.

(TS): I am not aware of hesitancy on the part of Discovery Service Providers to share search terms. Many EDS customers opt to integrate tools such as Google Analytics in their EDS profiles to track search terms and other usage behavior of interest.

The focus has been on journal article content, which is quite well established in discovery. Can you talk about the challenges of ebook content in terms of granularity? Do search results go to chapter/section which are highly desirable in making book cont

(TS): The level of granularity accessible in a Discovery Service is dependent on the level of granularity provided by publishers and content providers. Some content providers already support chapter-level ebook indexing, with chapter-level full-text PDFs and abstracts, so we already support chapter-level granular discovery in EDS. The issue is less a technical issue than a business model issue.

(DM): A lot of content providers are now shifting to chapter- and section- based delivery. A key challenge is that as a result of this shift there are many more individual items to serve, and each item needs to be individually (one-by-one) characterized/described with individual metadata.

With the proliferation of different metadata schemas, when will have a universal identifer for metadata at the chapter/article level?

(TS): I don’t know if we will ever have a universal identifier for this. We certainly could benefit from standardized tagging/coding of the granularity-level.

Where is EBSCO discovery getting the subscription information? Through aggregators or through direct subscription?

(TS): EDS customers manage their own subscriptions/holdings information in EBSCO’s knowledgebase product. We offer tools to help migrate holdings data from other non-EBSCO knowledgebase tools.

How do you relate parts to whole in your data schemes? How do you see search/discovery/delivery systems using those?

(DM): This is one of the areas where thesauri/taxonomies and ontologies provide a lot of value. Thanks to the hierarchical relationships that they include, it becomes easy to navigate between part and whole (or between topic and subtopic) in either direction. For example it becomes easy to specialize search only to a specific subtopic or to generalize it to the category without having to explicitly name all individual instances. Thanks to the adoption of semantic enrichment techniques, such navigation schemes are becoming more and more common in end-user interfaces.

Additional Information

  • Registration closes at 12:00 p.m. (ET) on March 11, 2015. Cancellations made by March 4, 2015 will receive a refund, less a $25 cancellation. After that date, there are no refunds.
  • Registrants will receive detailed instructions about accessing the webinar via e-mail the Monday prior to the event. (Anyone registering between Monday and the close of registration will receive the message shortly after the registration is received, within normal business hours.) Due to the widespread use of spam blockers, filters, out of office messages, etc., it is your responsibility to contact the NISO office if you do not receive login instructions before the start of the webinar.
  • If you have not received your Login Instruction email by 10 a.m. (ET) on the Tuesday before the webinar, at please contact the NISO office or email Juliana Wood, Educational Programs Manager at jwood@niso.org for immediate assistance.
  • Registration is per site (access for one computer) and includes access to the online recorded archive of the webinar. You may have as many people as you like from the registrant's organization view the webinar from that one connection. If you need additional connections, you will need to enter a separate registration for each connection needed.
  • If you are registering someone else from your organization, either use that person's e-mail address when registering or contact Juliana Wood to provide alternate contact information.
  • Library Standards Alliance (LSA) members receive one free webinar connection as part of their membership and DO NOT need to register for the event for this free connection. Your webinar contact will receive the login instructions the Monday before the event. You may have as many people as you like from the member's library view the webinar from that one connection. If you need additional connections beyond the free one, then you will need to enter a paid registration (at the member rate) for each additional connection required.
  • Webinar presentation slides and Q&A will be posted to the site following the live webinar.
  • Registrants and LSA member webinar contacts will receive an e-mail message containing access information to the archived webinar recording within 48 hours after the event. This recording access is only to be used by the registrant's or member's organization.