Data Infrastructure Initiatives Research Articles

How Data Infrastructure Deals with Bias Problems in Medical Imaging.

The paper discusses biases in medical imaging analysis, particularly focusing on the challenges posed by the development of machine learning algorithms and generative models. It introduces a taxonomy of bias problems and addresses them through a data infrastructure initiative: the PADME (Platform for Analytics and Distributed Machine-Learning for Enterprises), which is a part of the National Research Data Infrastructure for Personal Health Data (NFDI4Health) project. The PADME facilitates the structuring and sharing of health data while ensuring privacy and adherence to FAIR principles. The paper presents experimental results that show that generative methods can be effective in data augmentation. Complying with PADME infrastructure, this work proposes a solution framework to deal with bias in the different data stations and preserve privacy when transferring images. It highlights the importance of standardized data infrastructure in mitigating biases and promoting FAIR, reusable, and privacy-preserving research environments in healthcare.

Digitally Enlightened or Still in the Dark? Establishing a Sector-Wide Approach to Enhancing Data Synthesis and Research Potential in British Environmental Archaeology and Beyond

In a 2019 Internet Archaeology article, Elizabeth Pearson posed the question 'are we back in the Dark Ages?'. This question was made in reference to a developer-funded archaeology sector that was generating vast quantities of evidence and, particularly, in recent years, specialist environmental data, but was failing to mobilise this in a theoretical framework that generated meaningful advancement in terms of research. The introduction to the 2021 Internet Archaeology special issue on Digital Archiving in Archaeology (Richards et al. 2021) went on to address 'a digital resource that is now in jeopardy' – not only because of the risk of technical obsolescence, but also because of crucial limitations to its interoperability and discoverability. This article builds on these arguments and complements vital work underway on high-level, internationally focused data infrastructure initiatives (e.g. Wright and Richards 2018). We emphasise here the importance of parallel discussions at a community level, particularly with the people who routinely produce archaeological data, as key to enhancing data synthesis and research potential. Specifically, we report on two surveys conducted by the 'Rewilding' Later Prehistory project at Oxford Archaeology, in collaboration with Historic England and Bournemouth University, which originated in the 'Rewilding' project's concern with improving access to palaeoenvironmental data produced within Britain. Substantial amounts of zooarchaeological and archaeobotanical data remain buried in grey literature, limited-access publications and archive reports (not to mention floppy disks, CDs and microfiche), with no integrative means of searching for particular periods or categories of evidence. This lack of accessibility inhibits specialists from contextualising their findings, and was exemplified recently by the Archaeology on Furlough project tripling the known number of aurochs finds in Britain by trawling online records, journals and museum records (Wiseman 2020). The results of the surveys presented here, which targeted both environmental archaeologists specifically and the wider sector, demonstrate a significant appetite amongst archaeologists to improve data networks and for their work to contribute meaningfully to research agendas. Contextualised within a disciplinary landscape that is increasingly dynamic in its approach to tackling the openness and connectivity of 'big data', we argue that better data synthesis in environmental archaeology, and the developer-funded sector more broadly, can be more than just a mirage on the horizon, particularly once the people who produce the data are given an active voice in the matter.

Federated Heterogeneous Compute and Storage Infrastructure for the PUNCH4NFDI Consortium

PUNCH4NFDI, funded by the Germany Research Foundation initially for five years, is a diverse consortium of particle, astro-, astroparticle, hadron and nuclear physics embedded in the National Research Data Infrastructure initiative. In order to provide seamless and federated access to the huge variety of compute and storage systems provided by the participating communities covering their very diverse needs, the Compute4PUNCH and Storage4PUNCH concepts have been developed. Both concepts comprise state-of-the-art technologies such as a token-based AAI for standardized access to compute and storage resources. The community supplied heterogeneous HPC, HTC and Cloud compute resources are dynamically and transparently integrated into one federated HTCondorbased overlay batch system using the COBalD/TARDIS resource meta-scheduler. Traditional login nodes and a JupyterHub provide entry points into the entire landscape of available compute resources, while container technologies and the CERN Virtual Machine File System (CVMFS) ensure a scalable provisioning of community-specific software environments. In Storage4PUNCH, community supplied storage systems mainly based on dCache or XRootD technology are being federated in a common infrastructure employing methods that are well established in the wider HEP community. Furthermore existing technologies for caching as well as metadata handling are being evaluated with the aim for a deeper integration. The combined Compute4PUNCH and Storage4PUNCH environment will allow a large variety of researchers to carry out resource-demanding analysis tasks. In this contribution we will present the Compute4PUNCH and Storage4PUNCH concepts, the current status of the developments as well as first experiences with scientific applications being executed on the available prototypes.

Research Data Management and Data Sharing for Reproducible Research-Results of a Community Survey of the German National Research Data Infrastructure Initiative Neuroscience.

Science is changing: the volume and complexity of data are increasing, the number of studies is growing and the goal of achieving reproducible results requires new solutions for scientific data management. In the field of neuroscience, the German National Research Data Infrastructure (NFDI-Neuro) initiative aims to develop sustainable solutions for research data management (RDM). To obtain an understanding of the present RDM situation in the neuroscience community, NFDI-Neuro conducted a comprehensive survey among the neuroscience community. Here, we report and analyze the results of the survey. We focused the survey and our analysis on current needs, challenges, and opinions about RDM. The German neuroscience community perceives barriers with respect to RDM and data sharing mainly linked to (1) lack of data and metadata standards, (2) lack of community adopted provenance tracking methods, (3) lack of secure and privacy preserving research infrastructure for sensitive data, (4) lack of RDM literacy, and (5) lack of resources (time, personnel, money) for proper RDM. However, an overwhelming majority of community members (91%) indicated that they would be willing to share their data with other researchers and are interested to increase their RDM skills. Taking advantage of this willingness and overcoming the existing barriers requires the systematic development of standards, tools, and infrastructure, the provision of training, education, and support, as well as additional resources for RDM to the research community and a constant dialogue with relevant stakeholders including policy makers to leverage of a culture change through adapted incentivization and regulation.

The Federal Institute of Materials Research and Testing (BAM) – 150 Years of Enabling Scientific and Technological Breakthrough

The Federal Institute of Materials Research and Testing (BAM) – 150 Years of Enabling Scientific and Technological Breakthrough

Annotating OGC web feature services automatically for generating geospatial knowledge graphs

AbstractThe Spatial Data Infrastructure initiatives are now broadly developed and deployed. However, while plenty of tools use them, some tasks are still complex to perform by non‐expert users, such as finding, accessing, and using some of their related OGC Web Services (OWS). One of the main reasons for these challenges is associated with semantic heterogeneity within these services. This entails a lack of proper descriptions and requires knowledge about data structure and domain‐specific query languages to discover and retrieve these services. Semantic annotations of OWS play a crucial role in achieving semantic interoperability and addressing these associated challenges. In this article we describe an approach for enabling the automatic generation of semantic annotations of Web Feature Services (WFS) at their three request levels (GetCapabilities, DescribeFeatureType, and GetFeature), which are used to generate knowledge graphs. This approach uses various external services, ontological resources (POSM, DBpedia, GeoSPARQL, GeoNames, and datos.ign.es vocabularies), and knowledge bases (DBpedia and datos.ign.es). Moreover, this approach enables us to validate annotations obtained as a previous step to generating geospatial knowledge graphs. Additionally, we present our proposal through an application case and assess it using a representative set of 21 WFS services, achieving an average of 46.70% annotations, of which 22.29% and 35.52% were validated using DBpedia and datos.ign.es initiatives, respectively. This shows the feasibility of our approach to generate knowledge graphs from WFS services in an automatic way.

Research data management in clinical neuroscience: the national research data infrastructure initiative

Abstract In clinical neuroscience, there are considerable difficulties in translating basic research into clinical applications such as diagnostic tools or therapeutic interventions. This gap, known as the “valley of death,” was mainly attributed to the problem of “small numbers” in clinical neuroscience research, i.e. sample sizes that are too small (Hutson et al., 2017). As a possible solution, it has been repeatedly suggested to systematically manage research data to provide long-term storage, accessibility, and federate data. This goal is supported by a current call of the DFG for a national research data infrastructure (NFDI). This article will review current challenges and possible solutions specific to clinical neuroscience and discuss them in the context of other national and international health data initiatives. A successful NFDI consortium will help to overcome not only the “valley of death” but also promises a path to individualized medicine by enabling big data to produce generalizable results based on artificial intelligence and other methods.

Research Data in Chemistry – Results of the first NFDI4Chem Community Survey

Good research data management (RDM) requires a constant endeavour of the researchers but – first and foremost – user‐friendly infrastructure and well accepted standards need to be developed for the researchers. The national research data infrastructure initiative for chemistry in Germany NFDI4Chem sets out to embrace the challenge of creating a general research data management portal and hereby connecting already existing infrastructure as well as to foster the cultural change in chemistry towards digitalization by developing general minimum information standards for all methods used and teaching RDM principles to the community. In order to serve the needs of the chemical community at its best, NFDI4Chem accomplishes regular community surveys. The first survey has been performed in 2019 and the results were condensed into the NFDI4Chem proposal. This article summarizes the design of the study and its results. With regard to the project development, this first national survey serves as zero‐point of all upcoming efforts in research data management.

Transformative Use of an Improved All-Payer Hospital Discharge Data Infrastructure for Community-Based Participatory Research: A Sustainability Pathway.

To describe the use of a clinically enhanced maternal and child health (MCH) database to strengthen community-engaged research activities, and to support the sustainability of data infrastructure initiatives. Population-based, longitudinal database covering over 2.3 million mother-infant dyads during a 12-year period (1998-2009) in Florida. A community-based participatory research (CBPR) project in a socioeconomically disadvantaged community in central Tampa, Florida. Case study of the use of an enhanced state database for supporting CBPR activities. A federal data infrastructure award resulted in the creation of an MCH database in which over 92 percent of all birth certificate records for infants born between 1998 and 2009 were linked to maternal and infant hospital encounter-level data. The population-based, longitudinal database was used to supplement data collected from focus groups and community surveys with epidemiological and health care cost data on important MCH disparity issues in the target community. Data were used to facilitate a community-driven, decision-making process in which the most important priorities for intervention were identified. Integrating statewide all-payer, hospital-based databases into CBPR can empower underserved communities with a reliable source of health data, and it can promote the sustainability of newly developed data systems.

Judaica Europeana: An Infrastructure for Aggregating Jewish Content

Judaica Europeana envisions a world in which all digitized Jewish content in a variety of databases worldwide is aggregated and made accessible to users and applications anywhere, at any time. It seeks to set the ground so such content is cross-linked to conceptual structures (vocabularies, encyclopedias) that enrich them and provide contextual significance. Judaica Europeana is part of a cluster of projects building Europeana, a Linked Data infrastructure initiative of the European Commission. Judaica Europeana involves now some thirty-five partners from Europe, America and Israel, among them some of the most important Jewish content holders running long term digitization programs. It aggregated more than five million digital cultural objects and is continuing to process more. The data model (EDM) for describing these contents is that adopted by both leading world initiatives, Europeana and the Digital Public Library of America. The basic approach enabling EDM and based on the application of protocols and standards like RDF and Linked Data is surveyed and some actual examples of their current applications provided. The critical role of vocabularies for conceptual integration and access to contents is reviewed. A work program is outlined for the use of such vocabularies (thesauri, taxonomies, encyclopedias, etc.) to enrich the digitized content, interlink its diverse manifestations, and provide context and meaning. A first substantial achievement in carrying out such program is the publication of the YIVO Encyclopedia of Jews in Eastern Europe as Linked Data. Two main challenges facing the domain in the near future are detailed: (1) How to expand the availability, reaching a critical mass, of Jewish related vocabularies supporting queries like Who? What? When? Where? and expressed in the Linked Data/SKOS formats. (2) How the solid bases of such infrastructure so established may have an enabling effect in the development of new services: sophisticated offerings to the patrons of websites/portals, advanced K-12 ICT-based education, mobile cultural tourism applications, e-books, digital narratives storytelling, digital humanities scholarship, virtual research environments, MOOCs.

Virtual Australia: Developing an enabling platform to improve opportunities in the spatial information industry

The role that the development of Spatial Data Infrastructure initiatives is playing within the modern world is changing. Initially SDIs were implemented as a mechanism to facilitate access and sharing of spatial data hosted in distributed GISs. Users however now require precise spatial information in real time about real world objects and the ability to develop and implement cross‐jurisdictional and inter‐agency solutions to priorities such as emergency management, natural resource management, water rights, and animal, pest and disease control. In order to achieve this, the concept of an SDI is moving to a new business paradigm, where SDI is emerging as a ‘virtual jurisdiction’ or ‘virtual enterprise’ to promote the partnership of spatial information organizations (public/private) to provide access to a wider scope of data and services, of size and complexity that is beyond their individual capacity. The development of such an SDI requires an enabling platform to support the chaining of services across participating organisations. This paper outlines the outcomes of research on the development of such an enabling platform within the context of a Virtual Australia. This includes a review of the spatial information and SDI developments in selected Australian states (Victoria, Western Australia, Queensland and New South Wales) and international (Europe, Canada and USA) jurisdictions, which are helping to move the concept of an enabling platform forward, the improving opportunities for the spatial information industry through the development of an enabling platform, concepts and principles to aid in facilitating a Virtual Australia, and recommendations for the future.

Improving the rural data infrastructure of Scotland: an overview

Between 1945 and 1992 rural development policy in the UK was characterised by its sectoral approach, with agriculture dominating. Since then, reforms to the Common Agricultural Policy, and the UK government's commitment to sustainable development, have signalled a policy shift towards integrated rural development. Implementation is partly constrained by an institutional framework that evolved to set, deliver and evaluate sectoral policy. A particular legacy of this framework concerns the way data are gathered. This issue has been explored in a case study on the rural data infrastructure of Scotland. The findings highlight: the currently low level of user awareness of data sources; the lack of central government engagement in and individual awareness of related UK data infrastructure initiatives; and the current focus on top–down data provision being a constraint on community-led development. The commentary explores the interplay between policy and technology drivers and we argue that, whilst data sharing has been assumed to be a technical problem, there is evidence to suggest that it is more a social one concerning power relationships both within and between institutions. This is discussed in relation to other infrastructure projects in Europe and North America.

Improving the rural data infrastructure: the problem of addressable spatial units in a rural context

The transition from sectoral to integrated rural policy has highlighted the fact that data are highly institutionalised resources that are collected to satisfy the specific remit of the collecting organisation, often without reference to wider potential uses. This has led to a wide range of different spatial units being used for data collection, analysis and representation of data for the same location. These different spatial units are a significant barrier in creating a more integrated spatial data infrastructure to support Joined-Up Government. This paper uses Scotland as a case study to illustrate some of the barriers to integration. It has three aims. The first is to highlight the significance of these different spatial units in terms of their potential impact on statistical information derived by spatial overlay of data sets. The second is to review some of the potential solutions to defining consistent spatial data sets, as currently being developed through national spatial data infrastructure initiatives in the UK. The third is to identify the options for developing a system of addressable spatial units appropriate to sparsely populated rural areas.

Global Spatial data Infrastructure: Current Developments

A number of Spatial Data Infrastructure initiatives at local, national and European levels have been progressing since the early 1990s with a parallel global one emerging, beginning at the Bonn 1996 and continuing at the North Carolina 1997 meeting. This paper outlines some of the issues involved. Many of these are similar to those being experienced at the national and regional levels, but the solution to many of them will be different as will their implication. In particular it sets out the conclusions of the Chapel Hill, North Carolina meeting. The implementation of which is ongoing.

Building a North American Spatial Data Infrastructure

ABSTRACT: This paper addresses the state of spatial data infrastructures within North America in late 1997. After providing some background underlying the philosophy and development of the SDI concept, the authors discuss effects of technology, institutions, and standardization that confront the cohesive implementation of a common infrastructure today. The paper concludes with a comparative framework and specific examples of elements and initiatives defining respective spatial data infrastructure initiatives in the United States and Canada. Introduction Over the last three decades, governments and industry in countries around the world have invested tens of billions of dollars in the development of independent land systems designed largely to serve specific communities (forestry, land records management) within a local and regional framework. During the next decade the focus will increasingly shift to the challenges associated with integrating these systems together into--in each case--a national spatial data infrastructure (NSDI). This infrastructure will have the potential to act as an highway which links together environmental, socio-economic, and institutional databases (the horizontal highway) and provides for movement of from the local to national levels and, eventually, to the global community (the vertical highway). Proponents of the infrastructure concept often envision a data infrastructure as being analogous to networks of national highways (Branscomb 1982) or electric power grids in countries around the world. The highway metaphor conveys the notion of new or heightened connections to different locations. Further, as with electric power grids, the physical location of the source is usually less important to the end user than the continuing availability, reliability, and cost of the itself (Anderson 1990). These two metaphors have been widely employed over the past five years in defining and promoting a shared vision of the development of strategic networks and databases. Especially since 1990, a third metaphor--the information marketplace--has attracted the imagination and support of the business community (Coleman and McLaughlin 1998). The proposed development of national spatial data infrastructures has received considerable attention from institutional suppliers, the private sector, and user communities in developed countries around the world. Evolving from earlier data-sharing and program coordination efforts, the term NSDI has come to encompass the sources, systems, network linkages, standards, and institutional issues involved in delivering spatially related from many different sources to the widest possible group of potential users. The development of common frames of reference, technology, and related standards is crucial to the evolution of a data infrastructure that can support interoperability across provinces, states, and countries. This paper outlines the current state of corresponding spatial data infrastructure initiatives in North America and the issues affecting their implementation. From Integrated Mapping to Spatial Data Infrastructure The data-sharing paradigms underpinning the development of geo-spatial networks today are not new. In the 1960s, proponents of integrated mapping practices advocated the registration, overlay, interpretation and analysis of different layers or disparate themes of spatially related data sets to the practical solution of problems in land-use planning and inventory (McHarg 1969). Through the 1970s, the multipurpose cadastre concept launched major topographic and cadastral base-mapping megaprograms to support land administration at the local, state, and federal levels across North America, Australasia, and in emerging nations (McLaughlin 1975). By the early 1980s, the notion of information as a corporate resource (Diebold 1979) emerged, and the resources management movement encouraged individual organizations to implement enterprise approaches to the collection, management, and sharing of designated hard-copy and computer-based data holdings of corporate-wide interest. …