Socially-Derived Linking and Data Sharing within a Virtual Laboratory for the Humanities

HuNI (Humanities Networked Infrastructure) is a major new digital service for humanities researchers. Developed in Australia, with funding from the NeCTAR (National eResearch Collaborative Tools and Resources) programme, it aggregates data from 28 different cultural datasets from a variety of disciplines, makes them available for external re-use through an API and as Linked Open Data, and provides a set of tools for researchers to work with the data. HuNI is a virtual laboratory application which will be of great value to anyone interested in understanding Australia’s rich cultural heritage. It is also the single largest aggregation of networked humanities data in Australia – a new national data service which is of cultural significance in its own right, and accessible to all. HuNI is planned to go into production in the second quarter of 2014.

We reported on the initial design stage of this project at a previous Digital Humanities conference.1 This new paper will tell the full story of HuNI and will analyse, describe and evaluate its first full public release. It will include a demonstration of the full functionality of the service, and will report on its uptake by researchers and the wider community. We will also discuss the lessons learned from this large-scale project over its two-year lifespan, and the measures taken to ensure the sustainability of the HuNI service beyond the life of the project.

The paper will focus, in particular, on the ways in which HuNI is changing the nature of humanities research in the areas of data sharing, collaboration, community involvement, and the creation of socially linked data. Socially-derived linking of data is one of the key features of HuNI. Researchers are able to make assertions about relationships between entities represented in the aggregated data. If, for example, they search the data aggregate and identify two entities in their result set which are related in some way, they can add a link between the two records and define the nature of the relationship. The linking statement may be drawn from an existing vocabulary of relationships, or may use free text entered by the user. The virtual laboratory also allows a researcher to assert that two entities are not related, in recognition that this kind of statement is also a key characteristic of humanities research.

To help visualize these social links, each entity has its own network graph, showing up to six degrees of separation, resulting in a growing network of dynamic connections, or the “networked effect”. These “social linking” assertions are visible in the HuNI data aggregate. They may also appear in virtual collections assembled and published by individual users of the HuNi virtual laboratory. These virtual collections can be published for other users to see and re-use. Some links between entities have also been imported from the source datasets as part of the harvesting and ingest process. HuNI users can annotate these links with their own assertions as well.

Crucially, the provenance of all these “social linking” statements is also captured, enabling subsequent researchers to see who made each assertion. HuNI is an aggregate with a sense of its own history, in which researchers can trace how records have changed over time. Humanities research not only involves making connections between entities; it also involves assessing any changes in cultural flows and network relationships through time. So each HuNI record is time-stamped, meaning that researchers will always see the current view of a record, with its related records and assertions, but will also have the option to view how the record has changed since it was first harvested. The provenance information for each record, together with any curated assertions, is captured, so that researchers can see when the records were harvested and by whom. A link to the originating data record at source is also provided in the user interface.

“Social linking” is a crucial feature of the HuNI virtual laboratory. Instead of relying on a pre-coordinated mapping to a detailed ontology, we are relying on researchers and community users to establish most of the connections within the heterogeneous data aggregate. This enables HuNI to capture the different disciplinary perspectives of users, rather than trying to fit all the data into a single normative framework. It also acknowledges the productive differences that both define and link specific domains through a form of generative knowledge transfer. The opportunity to link data socially encourages HuNI users to share their knowledge and research findings in the form of specific assertions, and to discuss these statements with each other. In the paper, we will report on the ways in which this feature is being used by researchers and community users, and the extent to which it is enabling a new approach to data sharing in the humanities.

We acknowledge the contributions of Dr Marco La Rosa (Solution Architect) and Dr Anne Cregan (Semantic Lead and Business Analyst) in developing the HuNI infrastructure.

References
Burrows, T. (2012). Designing a national 'Virtual Laboratory' for the humanities: the Australian HuNI project. In Meyer, J. C. (ed), Digital Humanities 2012: Conference Abstracts. held July 16-22 at University of Hamburg. 139-141