The Transportation Research Thesaurus: Capabilities and Enhancements (2018)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

5 The Transportation Research Thesaurus (TRT) is a tool to improve the indexing and retrieval of transportation information. The TRT is used by the Transportation Research Board (TRB) and a variety of other organizations to support indexing, search, and retrieval of research documents, technical reports, and other information about transportation. The TRT also assists information users by facilitating consistency in language and cataloging information across repositories. At the same time, the TRT provides a basis for users who wish to tailor their own information storage and retrieval systems for their specific needs. While the TRT applies to “research” in particular, the TRT’s application is broader, cover- ing all modes and aspects of transportation and thereby providing a common and consistent language for communication among producers and users of transportation information. The TRT is curated and maintained by TRB with input from the TRT Subcommittee. The TRT web page—http://trt.trb.org/—allows for searching and browsing the thesaurus in various formats. The TRT originated as a tool to support TRB’s Transportation Research Information Ser- vices (TRIS) database, as described in NCHRP Report 450: Transportation Research Thesaurus and User’s Guide (Batty 2001), enabling indexers to describe documents in a consistent way and TRIS users to more easily retrieve relevant TRIS records in their areas of interest by searching the thesaurus terms. The TRIS database subsequently was consolidated with the Organization for Economic Cooperation and Development’s Joint Transport Research Centre’s International Transport Research Documentation (ITRD) Database into Transport Research International Documentation (TRID), the world’s largest and most comprehensive bibliographic source on transportation information. TRID contains more than one million records of published infor- mation and ongoing research, covering all modes and disciplines of transportation. The TRID database is available on TRB’s website at http://trid.trb.org. If the TRT is to maintain its value to a growing community of stakeholders, it must evolve to improve the user experience, make its services accessible to a broader range of users, accommo- date changing transportation technology, adapt to changing international information man- agement practices, and take advantage of advances in information technology. For example, the TRT’s functionality should ultimately facilitate (a) straightforward mapping between the TRT’s terms and those employed in the thesauri and other information management tools of other organizations; (b) implementation and maintenance of scalable and robust standards-based thesaurus management software (TMS); (c) an export function to enable users to adapt the TRT to meet their specific needs; and (d) use of the TRT by a wider range of producers, users, and curators of transportation information. Research is needed to develop a strategic plan for the TRT’s future development and to identify initial tasks that will immediately enhance the TRT and initiate implementation of C h a p t e r 1 Background

6 the transportation research thesaurus: Capabilities and enhancements the strategic plan. The strategic plan will direct the TRT’s future development to ensure that the thesaurus continues to serve well the needs of TRID users and the broader transportation community. Research Objectives The objectives of this research were to (a) develop a strategic plan for the TRT’s development and (b) identify a specific set of actions to enhance the TRT and initiate implementation of the strategic plan. Research Strategy The TRT of the future must be founded on a well-executed assessment of the current TRT and the transportation community’s vision of the role it plays in the future information envi- ronment. The assessment of current capabilities and potential enhancements included a critical review and assessment of the TRT, identification of opportunities for improvement, and pro- posed areas of enhancement. Conceptual Framework To organize the assessment, the research team developed a conceptual framework for a knowledge organization system that includes five dimensions, A through E. Figure 1-1 shows the dimensions, their detail, and interactions among them. Each dimension is further explained in the subsequent sections. Dimension A: Thesaurus Content, Management, and Maintenance The current version of the TRT includes terms, semantic relationships, ongoing day-to- day thesaurus management and maintenance tasks, and professional competencies and skills IM = information management Figure 1-1. Dimensions of a knowledge organization system.

Background 7 directly relevant to thesaurus management and information management. This dimension is the heart of the framework—without this component, there is no thesaurus. The TRT is a tool to consistently categorize research documents in TRB’s TRID database and assist TRID users in retrieving relevant information as described in NCHRP Report 450: TRT and User’s Guide (Batty 2001). Dimension B: Thesaurus Access and Use There are three categories of thesaurus access and use: explicit use (e.g., where the thesaurus is used directly by humans), use where it is embedded in another application (e.g., internal indexing, search system, recommender engine, dictionary or glossary term lookups from within a document, website, etc.), and custom use (e.g., based on specialized requests for use of parts or aspects of the thesaurus). Custom use designates products and services that may be generated to fulfill specialized requests. Each kind of access and use has implications for usability and for the architecture that supports the thesaurus. Failing to consider different kinds of access and use can lead to a suboptimal and constrained design, which, in the long run, may require fixes at additional cost or may prevent the thesaurus from supporting future opportunities. Dimension C: Thesaurus Governance Processes Standards in the field of information science are intended to provide guidance but will always be interpreted and adapted to the environment. Organizations can look to International Orga- nization for Standardization (ISO) 25964 only for general guidance on governance, and the governance process for the TRT should be designed to support the transportation community. Governance is all about the community and how the community works. Good governance is a balancing act between getting something done and ensuring that the full community is engaged. Dimension D: Governance Tools—Guidelines, Standards, Principles, and Best Practices Governance guidelines include all of those standards, principles, best practices, and warrants that are used in the governance and assessment processes. This dimension is important because at its foundation it guides the day-to-day management and maintenance of the thesaurus, guides design tasks, and ensures the trust and surety of the thesaurus. Dimension E: Thesaurus Architecture, Functionality, and Protocols The architecture and functionality of the thesaurus are important because they define what capabilities a user has to work with, what a user can and cannot do, and what products and services a user can produce. There are four levels of architecture to consider: (1) the architec- ture of the TMS application, (2) the architecture in which the thesaurus must function, (3) the communication architecture to support interoperability, and (4) the architecture required to support multilingual applications. Architecture Level 1. It is important to have a clearly defined data model for terms, their attributes and semantic relationships, and any elements needed to manage and organize the terms and vocabularies. For example, the current TRT defines facets through the extensive use of top terms and hierarchical relationships among terms. The current TRT is also sup- ported by an enumerated classification scheme, which provides a unique identifier for each

8 the transportation research thesaurus: Capabilities and enhancements term and controls the hierarchical display. This is traditional thesaurus architecture. For example: Water transportation (Aes) Shipping (Aesb) Ocean shipping (Aesbb) Coastwise shipping (Aesbc) Short sea shipping (Aesbe) In 2017, thesaurus architecture is usually grounded in a database structure where every term is represented by a full record, with administrative, access, and preservation fields. Architecture Level 1 also includes centralized versus federated support for thesaurus management functions, including permissions and version control. Additionally, Architecture Level 1 addresses issues of localization and collaboration. Some TMSs are designed to inherently support the creation and management of local vocabularies. These designs include all of the standard functionality for record locking and version control, as well as utilities for uploading, error checking, and reporting to manage exceptions. Note that the localization of vocabularies is different from micro-thesauri. Micro-thesauri are part of a larger macro-thesaurus structure and are created and maintained through the use of categorization or classification schemes (Cambridge System- atics, Inc. 2013). Architecture Level 2. The second kind of architecture depends on how the thesaurus is used within the organization and the level and nature of integration with other applications. This architecture may take the form of tight integration or loose coupling, depending largely on the type and extent of consuming applications. Generally, business rules for use and implementation of the thesaurus are addressed at this level. Architecture Level 3. This level is generally defined by communication protocols such as the Simple Knowledge Organization System (SKOS) or Resource Description Framework (RDF) that allow an organization to identify and define the structure and elements of their knowledge organization system for consumption by other organizations. The SKOS may also be used as an output format for consumption by other applications. The SKOS, though, should be defined to support the data model of the thesaurus. Architecture Level 1 of the thesaurus should not be defined by the SKOS, but Architecture Level 3 should have the capability to export the thesaurus or any part of it in SKOS format. Architecture Level 4. Architecture Level 4 pertains to the structures required to support multilingual content and applications. The requirements for displaying records in a multilin- gual thesaurus are different from those that are required to support multilingual search and machine translation. This conceptual framework identifies at a high level the key challenges and opportunities that are important to delivering a successful result for the TRT community. Because the TRT has many stakeholders who play a variety of roles, the conceptual framework also provides a holistic and multifaceted view of the various dimensions of the thesaurus. A common understanding across the community is essential for making practical and wise decisions about the future of the TRT. Assessment Steps With this framework as background, the assessment of the current TRT was conducted in three steps:

Background 9 • Step 1: Formulate assessment criteria in consultation with the project panel and the trans- portation community, • Step 2: Assess the current state of the TRT against assessment criteria, and • Step 3: Identify gaps and opportunities for improvement and enhancement in the con- text of the TRT vision. Each step in this critical task had its own detailed methodology. Once completed, these steps were synthesized and structured into a draft report for review by the project panel. The three steps are discussed below. Step 1: Formulate Assessment Criteria In order to provide a foundation for understanding the current state of the TRT, the research team identified assessment criteria through a review of the peer-reviewed literature, the gray literature, and in consultation with the research panel members. The research team then aligned each criterion with a dimension of the conceptual framework. In addition, the criteria were grouped by impact categories within each dimension to facilitate assessment. The research team identified 166 criteria across the dimensions. The full list of criteria organized by dimension is provided in Appendix C. Step 2: Assess the Current State of the TRT Against Assessment Criteria The assessment against the criteria developed in Step 1 was completed through two activi- ties. The first activity focused on a detailed description of the TRT based on the framework and criteria. The research team realized early in the process that a consistent, complete, and comprehensive understanding of the TRT was important as a foundation, prior to undertak- ing a critical assessment. This description represents a non-critical, fact-finding approach. The second activity involved the critical assessment of the current state of the TRT. The assessment was grounded in the criteria-based description from Step 1. The critical assessment, presented in Chapter 3, discusses overall strengths and weaknesses of the TRT, based on each dimension of the conceptual framework. Step 3: Identify Gaps and Opportunities for Improvement and Enhancement Unlike the description and critical assessment of the current state of the TRT, which were straightforward tasks grounded in data and information, defining future opportunities pre- sented challenges. The opportunities must be practical and affordable and be designed in such a way to avoid loss of functionality or disruption to existing applications. The opportunities must build value and capacity incrementally and be achievable over time. These opportunities must also be designed to accommodate future changes in the larger transportation field and in information and semantic technologies. The decision points and alternatives must be clearly specified for those opportunities. To meet this challenge, the research team identified five use case scenarios. The five use cases were defined to progressively add value over time and to provide decision makers with options for defining the TRT’s future. Use Case 0 represents the status quo. Use Case 1 focuses on the creation of a new master version of the TRT, which is standards based and will expand engagement with and use by the transportation community. Use Case 2 further develops the new master version of the TRT as an important information discovery tool for the field of transportation. Use Case 3 explains how the TRT could be developed to support both centralized and local automated categorization and indexing. Finally, Use Case 4 describes a designer version of the TRT that serves as the foundation for future tools and capabilities,

10 the transportation research thesaurus: Capabilities and enhancements including those developed by the broader transportation community. The use case scenarios are presented in more detail as future strategies in Chapter 4. Sources of Data and Information for the Assessment The research team drew upon three major sources of data and information to complete the assessment: (1) a working copy of the TRT, (2) interviews and online data collection from rep- resentatives of TRT stakeholder groups, and (3) focused searching of comparable transporta- tion information sources. In addition, the assessment, findings, and conclusions were informed by a detailed analysis of the TRT Facet X (Information Organization), from development of 50 definitions for terms already in the TRT, and from a market assessment of TMS. These “deep dive” tasks provided an opportunity for the research team to assess and evaluate aspects of the TRT in more detail. Working Copy of the TRT The research team began the project with access to the current web version of the TRT, all of the supporting documentation created by the TRT management team, and access to the admin- istrative interface of the TRT. While these were rich sources of information for the assessment, the research team needed a standard thesaurus management and reporting tool in order to complete the assessment. For this reason, the research team set out to generate a working copy of the TRT in a thesaurus management tool. Having a full and complete copy of the thesaurus was necessary to conduct the assessment and provide all stakeholders with an accurate picture of both the TRT’s current state and possible future strategies. The TRT team provided several sources for the research team to work with to create the work- ing copy. However, several challenges were encountered in the generation of this working copy, and it became clear that the only way to gain an accurate representation of all TRT relationships was to capture and reconstruct full records at the individual term level. Figure 1-2 presents a snapshot of a full term TRT record reconstructed and transformed into a record compliant with American National Standards Institute (ANSI)/National Information Standards Organization (NISO) Z39.19 and ISO 25964. It includes the TRT components that serve as the source for the reconstructed and transformed record. With a fully ISO-compliant thesaurus available in an application with extensive search and reporting capabilities, the research team was able to generate the data needed to address a large number of the assessment criteria. Stakeholder Interviews and Feedback To gain a better understanding of the current TRT environment, as well as challenges and opportunities, the research team conducted interviews with five stakeholder groups: (1) depart- ment of transportation (DOT) staff who have critical roles pertinent to the TRT; (2) transporta- tion subject matter experts (SMEs) who are the primary intended users of the TRT; (3) the TRT administrative staff, TRB project panel and the TRT Subcommittee; (4) the TRT system architects and administrators; and (5) information science professionals in state transportation department libraries, academic transportation libraries, and transportation knowledge networks. The research team used its network of contacts and input from the project panel, subcommit- tee and TRT staff to identify representatives for these five groups. The SMEs were selected based on expertise in each of the 21 high-level facets of the TRT.

Background 11 An interview guide (see Appendix D) was developed and distributed prior to each interview. This guide included a general section about the purpose and protocol for conducting the inter- views, followed by specific sets of questions for each interview group. In some cases, a particular individual represented multiple groups, in which case the question sets were combined. The questions served as a preliminary set of questions to guide the interviews. For each interview, the research team recorded information in a distinct document and, with permission, recorded the interview, if feasible. Each document was filed and archived as part of the research effort. In addition, reference materials shared by the interviewee were logged and filed in the research archive. The research team conducted a total of 26 interviews, including five members of the NCHRP Project 20-109 Panel and key DOT stakeholders, nine state DOT information professionals and academic transportation librarians, eleven SMEs, and the TRT system administrator. The infor- mation gathered through interviews supplemented the quantitative assessment data generated by the research team working directly with the TRT. The research team conducted hour-long, virtual interviews with several state and academic transportation librarians and managers. Several attempts were made to contact librarians in other states. When these states were unresponsive to the invitation, the research team shifted its strategy to an online data collection form (see Appendix D). The online data collection form was distributed to 81 state DOT, academic, and private organization trans- portation librarians. Fourteen responses were collected. These responses provided a level of feedback the research team hoped to achieve through the interactive interviews. This approach was also used to collect additional information from SMEs. Sections 6 and 8 of the Figure 1-2. TRT hierarchical structure compared to ISO 25964 structure.

12 the transportation research thesaurus: Capabilities and enhancements in-person interview guide for SMEs were transformed into an online data collection form (see Appendix D). Focused Searching of Peer Transportation Information Sources A significant number of the criteria in Dimension D (Governance Tools—Guidelines, Stan- dards, Principles, and Best Practices) deal with the alignment of the TRT with the scope and coverage of the field of transportation. In addition to the transportation research found in the National Academies of Sciences, Engineering, and Medicine information repositories, trans- portation research is published in peer-reviewed journals, technical reports, and gray literature, accessible through open source search systems such as Google Scholar, as well as in commercial databases. The research team conducted two types of tests to gauge the alignment of the TRT with the field. The first type was a general search of terms in glossaries, manuals, and encyclo- pedias referenced by the research team’s Advisory Panel of SMEs for use in the TRT assessment. The second type of search involved searching the TRT terms in other transportation databases and publications to determine their productivity and coverage. Review of Previous NCHRP Reports In addition to interviews and focused searching for terms, the team reviewed previous NCHRP reports that included assessments of or references to the TRT. All of these earlier proj- ects and reports were leveraged in this research, which is the first comprehensive assessment of the TRT. The TRT was initially developed under NCHRP Project 20-32 (2001) to provide a tool to improve the indexing and retrieval of transportation information. The project produced a machine application to replace the previous paper version of the TRT, updated the structure and content, and created a user’s guide to support the implementation and use of the TRT. The ini- tial project leveraged the Viewer software developed by CDB Enterprises in 1994/1995. While the original design of the TRT was intended to be ANSI/NISO Z39.19 compliant, the use of the Viewer software and its limited application as a component of search resulted in a suboptimal thesaurus application. In 2007, NCHRP Project 20-70 was undertaken to develop a database at TRB that would allow TRB to maintain the authoritative version of the TRT in-house (i.e., to eliminate the external contractor role), to integrate the TRT with the then-new TRIS system for creating catalog records, and to publish the TRT on the Internet. Seven specific tasks were identified and completed. This project resulted in the current version of the TRT (http://trt.trb.org/trt.asp?). When NCHRP Project 20-70 was launched, the architectural design was guided by the TRIS database. Alignment with ANSI/NISO Z39.19 and/or ISO 25964 standards was not undertaken. These standards were not provided to the contractor as design specifications. In 2008, NCHRP Project 20-79 was launched. This project focused on the addition of AASHTO definitions to the TRT and was completed in 2009. NCHRP Project 20-79 was imple- mented in the digital version of the TRT, a component of TRIS. While the project resulted in the integration of some existing definitions, a comprehensive strategy for the use and inclusion of definitions was not developed. Since 2009, two additional NCHRP research projects have addressed or mentioned the TRT, NCHRP Project 20-90 and NCHRP Project 20-97. NCHRP Project 20-90 was launched in 2011 and completed in 2013 by Cambridge Systematics. The objective of this research was to prepare guidance describing the practices that state DOTs could use for capture, preservation, search,

Background 13 retrieval, and governance of transportation data and information, as well as strategies and actions a DOT could follow to implement such practices. As part of this broader project, there was a small targeted review of TRT as a cataloging tool. While some suggestions were offered to enhance the TRT, the goal was not to provide a comprehensive review of the thesaurus. NCHRP 20-97 referenced the TRT as a tool that was designed to support findability of transpor- tation information. The goal of this project, though, was to improve the findability of state DOT information by defining an information management framework, describing good practices for organizing and classifying information, and developing federated search procedures. The TRT was mentioned but was not treated or leveraged in depth for this research project. Since 2011, the state DOTs have also provided feedback and comments on the TRT, in particular feedback pertaining to the scope and coverage of the TRT and the general usability and ease of access to the resource.