Introduction
Many VIVO implementers find collecting, mapping, and loading data into VIVO to be quite difficult. For example, data on publications, grants, and datasets produced by an institution’s faculty can be difficult to find and disambiguate. Understanding the ontologies used to describe data in VIVO and mapping faculty data to those ontologies involves a steep learning curve. Also, transforming the data to a linked data format, such as VIVO RDF, has proven difficult for most implementers due to gaps in skills and knowledge. These barriers have prevented organizations from joining the VIVO community and adopting the technology that enables access, discovery, and analysis of scholarship data.
...
An Expression of Interest (EOI) notice soliciting participation in a Research Graph VIVO Cloud Pilot was distributed at the Open Repositories conference in June 2017 as well as at the eResearch conference on October 2017, both located in Australia. The notice was also distributed to peers in Germany and Canada. Four organizations formally expressed interest in participation. Another five organizations informally expressed interest in participating..
Pilot Phases and Structure
The organizations that expressed interest in the Research Graph VIVO Cloud Pilot have varying levels of knowledge of the projects and technologies. The first phase will enable the Cloud Pilot Team to confirm what they consider to be the most important and unknown technical variable of the project, e.g. how large the seed data will grow after first and second level connections are made and what impact that will have on hosted server resources, cost, and performance.
The Cloud Pilot will also provide excellent information on scaling the service. Using subsets of Pilot Organization data, we can identify small to medium graph sizes. Using their entire faculty as seed data, we can load and scale test to large sizes. The organizations identified for the first phase of the Cloud Pilot are thought leaders in linked data, repositories, and the open science community. Their involvement lowers project risks and helps build understanding of a potential international service offering.
Pilot Assumptions and Risks
- Market analysis -- We assume that there is a market for the production of VIVO data and turn-key hosted VIVO sites at a reasonable price. However, the market analysis may indicate there is no market for the service
- Technical effort -- Duraspace is currently gapped with respect to VIVO technical knowledge due to staff turnover. We assume this gap can be quickly filled by existing Duraspace staff, assisted by Dr. Conlon
- Customization -- we assume that the customer can be satisfied with simple theming (colors, logo, site name) of the turn-key hosted site, and that the theming can be delivered at reasonable cost
- Graph size -- the number of entities and triples -- resulting from particular seed datasets is not well understood. Based on the experience of large VIVO sites (Duke, Vidwan, Florida), we do not expect this to be an issue, even for large seed datasets (10,000 researchers). We propose a limit on the size of the final graph to be 500,000 entities for the purposes of this Cloud Pilot
- Data value -- the data produced by Research Graph must be of high value to the customer, including significant coverage and accuracy of first and second order entities
- Team member commitment -- we assume that the Pilot Team members (see below) and pilot organizations can provide the required effort in the required timeframe to participate in a Cloud Pilot Working Group
Proposed Pilot Timeline and Effort
The following effort is estimated for the Pilot Team during the term: Market analysis and service definition lead (10%), Project manager (10%), VIVO subject matter expert (10%), Duraspace technical resource (20%), Research Graph subject matter expert (5%), and a Research Graph technical resource (20%). We recommend The Pilot Team meet with pilot organizations weekly during the term, forming the Cloud Pilot Working Group.
...