The EarthCollab project, officially titled "EarthCube Building Blocks: Collaborative Proposal: Enabling Scientific Collaboration and Discovery through Semantic Connections," is focused on using Semantic Web and linked data technology to facilitate the coordination and organization of complex scientific projects and their products. The premise of EarthCollab is that many relationships exist among (and within) scientific resources, projects, and people and organizations, and these relationships provide an important means for people to discover and use geoscience data resources. From a technology point of view, the project has developed information systems that leverage linked data to produce more coherent methods of information and data discovery for large multi-disciplinary projects and virtual organizations. These systems enable researchers to more easily find people, organizations, and research resources that are relevant to their work, which could spark new scientific discoveries.

The project has used the open source VIVO semantic software suite (http://vivoweb.org) to provide interfaces for researchers within two targeted scientific research areas to explore the people, publications, platforms, and data sets within their respective communities. These systems, called "Connect UNAVCO" (https://connect.unavco.org/) and "Arctic Data Connects"(http://vivo.eol.ucar.edu), are both live for public use. As an illustration of the kinds of information that these systems contain, as of 31 August 2018, Connect UNAVCO contains records for nearly 6,000 scientific documents, 4,620 datasets, 3,920 research sites, 236 grants, 846 people, and 381 organizations. Arctic Data Connects, with a more focused case study on select arctic research projects, contains records for 354 datasets, 26 grants, 146 people, and 53 organizations. We used multiple user-centered design methodologies to inform the development and iteration of Connect UNAVCO and  Arctic Data Connects, including targeted surveys, small focus groups, and individual interviews and usability tests.

The information in these systems comes from a combination of existing metadata databases and newly created metadata. The resources are represented via the Semantic Web data model, in which ontologies are used to define entity types (classes), and the relationships between them (properties). To facilitate easier integration and data sharing across geoscience communities, our goal within this project has been to reuse existing ontologies as much as possible when developing project ontologies and web applications. This effort involved conceptual modeling of the key entities and relationships of interest, and finding relevant existing ontologies that map to our conceptual model. To create the EarthCollab ontology model, we combined components from multiple ontologies, and created custom ontology extensions as necessary to fill gaps.

The result is networked sets of information, in which researchers, datasets, publications, scientific instruments, research projects, and research locations are each represented in the data model as distinct entities that have specific types of relationships with other entities. Practically, the VIVO software displays a web page for each entity, which provides information about the entity along with links to other entities that are related via explicitly declared relationships. For example, the web page for a particular dataset within one of the systems will potentially display links to associated publications, organizations, grants, creators, and instruments.

Another component of this project has been to add capabilities to the VIVO software suite itself. In the first two years of the project, we developed a prototype "cross-linking" approach to exchange information across VIVO instances. The motivation for this work was that scientific research is increasingly multi-organizational. The new VIVO cross-linking feature was developed to enable the exchange of information about specific people or entities that are in common across different VIVO instances. This new feature reduces duplication of information across systems and enables the distribution of authoritative information about specific entities, such a single individual who has VIVO profiles at different organizations. The cross-linking capability has been deployed to operational VIVO systems, and is being contributed to the VIVO open source code community.

In summary, some of the advances of the project include further development and enhancement of the open source VIVO software for sharing and discovering research; development of VIVO instances at two university consortia (UNAVCO and NCAR EOL) with the unique challenges of extending VIVO beyond a single organization use case; consideration, discussion, and advancement of geoscience semantics to improve connectivity of research outputs; consideration and extension of appropriate ontologies into the geoscience research space; advancing the use of persistent identifiers for researchers, publications, datasets, data products, and institutions to improve research discovery; significantly enhancing the connectivity of people, publications, and products within the two consortia; consulting, advising, and informing researchers about ways to connect and discover research; and most importantly building two significant VIVO instances, Connect UNAVCO and Arctic Data Connects, to enhance sharing and discovery.

Last Modified: 10/29/2018
Modified by: Dean Krafft