Taxonomy Sessions at the 2019 SLA Conference

SLA (Special Libraries Association) offered a good number of taxonomy-related sessions at this year’s annual conference, held June 14-18 in Cleveland, Ohio, thanks to the organizing efforts of its Taxonomy Division. There were enough taxonomy sessions so that there was always at least one session of interest at any time.

SLA is a membership association of librarians and information professionals, particularly involved in “special” libraries or information services. Special libraries include corporate, specialized academic, government, military, law, medical, business, and nonprofit. I’m not a librarian (I’m an accidental taxonomist), so I didn't become a member of this professional organization until a Taxonomy Division was created 10 years ago. I’ve attended and presented at some, but not all of the SLA conferences in the past 10 years, as the taxonomy-related offerings vary, and presentation topics are usually the choice of the Taxonomy Division conference planning committee. This year, for the first time, I presented not one, but two sessions: the full-day preconference “continuing education” workshop and co-presented a session on taxonomy management software.

I was very pleased that there was such a rich program in other taxonomy sessions this year, especially compared to last year, thanks to Taxonomy Division conference program chair Janice Keeler and program committee members Edee Edwards and Margaret Nunez. There were also two sessions on knowledge management, which I found very interesting. The taxonomy-related sessions were:

• "Introduction to Taxonomy Design & Creation" (full-day preconference workshop)
• "Ensuring Semantic Interoperability and Creating Interoperable Taxonomies" (90 minutes, 2 speakers)
• "Taxonomy Governance in Real Life" (90 minutes, panel discussion, 2 speakers and moderator)
• "Taxonomy Roundtable” (90 minutes, three roundtables of participant discussions)
• "Big Data and Controlled Vocabularies (30 minutes, 2 speakers)
• "Taxonomy Basics" (30 Minutes)
• "Keeping your Taxonomy Fresh and Relevant" (60 minutes, 2 speakers)
• "Taxonomy-Ontology Conversions: Case Studies (75 minutes, 3 speakers)
• "Taxonomy Tools and Tool Evaluation" (60 minutes, 2 speakers)

“Ensuring Semantic Interoperability & Creating Interoperable Taxonomies,” was a densely packed presentation covering the different types and issues in controlled vocabulary interoperability with two presenters in turn: Margie Hlava, President and Chairman, Access Innovations Inc., and Marcia Zeng, Professor of Library and Information Science, Kent State University. Marcia explained that interoperability is at different levels: system level, semantic level, and structural level, and her focus was on semantic interoperability, which is addressed in the international standard ISO 25964-2. She discussed in detail each of the different kinds controlled vocabulary interoperability. There are methods that are based on working from an existing knowledge organization system: derivation from an original source and expansion; and there are methods that involve working between/among existing vocabularies: integration/combination and interoperation/shared/harmonization.

“Taxonomy Governance in Real Life” featured two panelists of very different organizations: Paula McCoy, manager from ProQuest, and Susannah Woodbury, taxonomist from Overstock. Taxonomy governance was defined as maintaining the content of a controlled vocabulary (adds/deletes), maintaining the integrity of a vocabulary (standards and usage), and implementing a vocabulary (managing those who work in and use the vocabulary). Topics of discussion included working with stakeholders, the governance of change and how decisions are made, and staying flexible through the iterative process.

The “Taxonomy Roundtable” was a purely discussion-based session, whereby attendees divided into three groups of about 6-7, and each group got to discuss three of the four predefined topics in turn: taxonomy ROI, adding taxonomy to the workflow, implementing taxonomies in search, and taxonomies in user interface design. These topics were chosen based on a Taxonomy Division survey of members’ interests. Each table then reported the outcomes of their discussions to the larger group.

“Big Data and Controlled Vocabularies” was presented by Camille Matthew, Information Science Specialist, NASA Jet Propulsion Laboratory. Camille explained what big data is: an accumulation of data that is too large and complex for processing by traditional database management tools. Big data is big by 5 Vs: volume, velocity, variety, veracity (varied dates/outdated for example), and value. The issue is combining “stuff” (big data) and “strategy.” Strategies include controlled vocabulary, taxonomy, ontology, and metadata standards. Data structure cannot be assumed, so we design for unstructured content.

"Taxonomy Basics" was presented by Heather Kotula, Vice President of Marketing and Communications, Access Innovations Inc. This quick session was aimed at those new to taxonomies. It comprised definitions of taxonomies and other types of controlled vocabularies and also included quite a bit of history into the field of classification and naming.

“Keeping your Taxonomy Fresh and Relevant: The APA Thesaurus” was presented by Marisa Hughes, Taxonomist, American Psychological Association (APA). Marisa had recently led a thorough thesaurus update that took about a year and was completed in February 2019, and this presentation was largely based on lessons learned from that project. Change management is a key part of in taxonomy governance. Change is constant. Creating a responsive and relevant taxonomy involves a set of activities: adapt, determine, engage, delineate, data, identify. One needs to know when and why to change, and a roadmap is also needed.

 “Taxonomy-Ontology Conversions: Case Studies” comprised three case-study presenters: Edee Edwards, Ontology Architect at the National Fire Protection Association (NFPA); Mary Chitty, Library Director & Taxonomist at Cambridge HealthTech; and David Bender, Manager, Medical Ontology, Radiological Society of North America. The genesis of this session came out of a conversation at the SLA conference the previous year, when someone from Lexis Nexis asked you build your ontologies: from taxonomies or from data. It was anticipated that the case studies would be all be conversions from taxonomies to ontologies, but that was not necessarily so.

Edee Edwards explained that the NFPA was building a data science team, which was very interested in ontologies. There was also a data governance group involved. At that time, they also needed a system upgrade of vocabulary software, and the new one is SKOS-based. They did a proof of concept with our data science group. NFPA’s primary use for the ontology was auto-tagging.  

Mary Chitty's presentation “Preparing your taxonomy to be ready for data scientists & machine readability " presented the case of taxonomies at Cambridge HealthTech. It is still a taxonomy, not an ontology, but Cambridge HealthTech has recently partnered with OntoForce, a semantic search and data science company to use their search engine. The presentation was more about the issues than any solution. Ongoing challenges include dealing with legacy data, integrating acquired companies’ data, scaling up, and dealing with ambiguity.

David Bender’s presentation “The Big Maybe: Should You Convert Your Taxonomy to an Ontology?” presented the example of the controlled vocabulary of the Radiological Society of North America (RSNA), RadLex. RadLex, a model of anatomical procedure and modality as it pertains to radiology, is referred to as a lexicon or terminology, although it is arranged as a hierarchical tree/taxonomy. It was decided to have a structure, as an ontology, but there are still more questions than answers. In moving in the direction of ontology, they are using the tool Protégé and have converted RadLex into an OWL form, but otherwise it is still kept as a taxonomy.

I presented the full-day preconference workshop "Introduction to Taxonomy Design & Creation"  and on taxonomy management software in the session "Taxonomy Tools and Tool Evaluation, and my co-presenter Marti Heyman of OCLC presented on how to evaluate taxonomy tools.  

SLA Taxonomy Division members can read detailed reports of each of these taxonomy sessions in the next issue of the Division’s “Taxonomy Times” newsletter.

SLA, an international organization, holds its annual conference in June in different cities in North America. It will next be in Charlotte, North Carolina, June 6-9, 2020.