Towards a Unified University Infrastructure: The Data Management Roll-Out at the University of Oxford

Abstract

Since presenting a paper at the International Digital Curation Conference 2010 conference entitled ‘An Institutional Approach to Developing Research Data Management Infrastructure’, the University of Oxford has come a long way in developing research data management (RDM) policy, tools and training to address the various phases of the research data lifecycle. Work has now begun on integrating these various elements into a unified infrastructure for the whole university, under the aegis of the Data Management Roll-out at Oxford (Damaro) Project.This paper will explain the process and motivation behind the project, and describes our vision for the future. It will also introduce the new tools and processes created by the university to tie the individual RDM components together. Chief among these is the ‘DataFinder’ – a hierarchically-structured metadata cataloguing system which will enable researchers to search for and locate research datasets hosted in a variety of different datastores from institutional repositories, through Web 2 services, to filing cabinets standing in department offices. DataFinder will be able to pull and associate research metadata from research information databases and data management plans, and is intended to be CERIF compatible. DataFinder is being designed so that it can be deployed at different levels within different contexts, with higher-level instances harvesting information from lower-level instances enabling, for example, an academic department to deploy one instance of DataFinder, which can then be harvested by another at an institutional level, which can then in turn be harvested by another at a national level.The paper will also consider the requirements of embedding tools and training within an institution and address the difficulties of ensuring the sustainability of an RDM infrastructure at a time when funding for such endeavours is limited. Our research shows that researchers (and indeed departments) are at present not exposed to the true costs of their (often suboptimal) data management solutions, whereas when data management services are centrally provided the full costs are visible and off-putting. There is, therefore, the need to sell the benefits of centrally-provided infrastructure to researchers. Furthermore, there is a distinction between training and services that can be most effectively provided at the institutional level, and those which need to be provided at the divisional or departmental level in order to be relevant and applicable to researchers. This is being addressed in principle by Oxford’s research data management policy, and in practice by the planning and piloting aspects of the Damaro Project.

Digital curation and costs

“…However, these also bear costs regarding storage and data preparation for curation [55]. Rice et al [43] argue that it is the duty of research institutions to develop infrastructures that support data variety and their reuse, especially when considering that, according to Wilson & Jeffreys [56] it is the case of research data supported by a public budget. For this reason, Delasalle [16] points out that one of the challenges is to overcome skepticism amongst researchers.…”

“…et al [8] stand against a wide data dissemination because it is not free, which instead implies significant costs identified by Whyte & Pryor [55] as relating to data understandability (preparation and documentation using acknowledged meta-information schemes -revision costs) to conclude that sharing should increase the cost, instead of decreasing it. Dürr et al [20] recommend research on displaying tools and on recovery of data sets, whereas Wilson & Jeffreys [56] advocate costs in disseminating data management services. h) Administration: The positing of monitoring, as expressed by Kejser,et al [30] as the vigilance of the target community (costs inversely vary according to the influence of the curating body in producing and using formats) and technology (costs depend on development and complexity).…”

“…o) Costs: Studies identify fixed and variable direct costs and indirect costs. Wilson & Jeffreys [56] note that institutional services often demonstrate high fixed costs when comparing them with variable costs (the urge for constant development and human resources), while Whyte & Pryor [56] argue that it is necessary to efficiently use short resources for data collection -including research topics and instrumentation -in order to reduce indirect costs. Consequently, Kejser et al [30] express the idea that models should represent complete economical costs, whether direct or indirect.…”

“…Additionally, despite the funder's attempt to identify sustainable resources, they often neglect the creation of a clause to measure whether or not the agreed results are accomplished or measured [39]. In one of the studies about management and business models, Wilson & Jeffreys [56] conclude that business models have a better production at the level of individual components, rather than at the level of the whole infrastructure. Poole [39] argues that the development of business models, as well as cooperation through preservation networks are essential concerns when computing costs, and are deeply related to the issue of costs sharing, as exemplified by Suchodoletz et al [51].…”

“…Queiroz [41] considers it an activity that aims to obtain best practices regarding the processing of digital objects, depending on production and technological contexts affected by human, socioeconomic, financial variables, amongst others. Bicarregui et al [8] approach to curation underlies management's planning and data preservation for the Big Science for on-going use, and Wilson & Jeffreys [56] argue that specialized repositories possess an acknowledged experience in the curation skills of their human resources, while Dillon [17] and Strasser et al [50] highlight the need for researchers to receive training in data management and digital curation in order to acquire good practices. From an applied linguistics and journalism perspective, Barros [6] states that digital curation is related to contents' selection, organization, understandability and evolution, representing a new practice that media professionals and volunteering users have to partake in, and that can only be understood when considering its cultural framework.…”