Search

In 2020, the COVID-19 pandemic resulted in a rapid response from governments and researchers worldwide, but information-sharing mechanisms were variable, and many early efforts were insufficient for the purpose. We interviewed fifteen data professionals located around the world, working with COVID-19-relevant data types in semi-structured interviews. Interviews covered both challenges and positive experiences with data in multiple domains and formats, including medical records, social deprivation, hospital bed capacity, and mobility data. We analyze this qualitative corpus of experiences for content and themes and identify four sequential barriers a researcher may encounter. These are: (1) Knowing data exists, (2) being able to access that data, (3) data quality, and (4) ability to share data onwards. A fifth barrier, (5) human throughput capacity, is present throughout all four stages. Examples of these barriers range from challenges faced by single individuals to non-existent records of historic mingling/social distance laws, and up to systemic geopolitical data suppression. Finally, we recommend that governments and local authorities explicitly create machine-readable temporal “law as code” for changes in laws such as mobility/mingling laws and changes in geographical regions.

Abstract Openness and transparency in the research process are a prerequisite to the production of high quality research outputs. Efforts to maximise these features have substantially accelerated in recent years, placing open and transparent research practices at the forefront of funding and related priorities, and encouraging investment in resources and infrastructure to enable such practices. Despite these efforts, there has been no systematic documentation of current practices, infrastructure, or training and resources that support open and transparent research in the UK. To address this gap, we developed and conducted the Open and Transparent Research Practices survey, a large-scale audit study completed by research-active staff in UK research institutions to better understand existing practices, needs, support, and barriers faced when implementing open and transparent research. The data presented here capture responses from over 2,500 research-active staff based at 15 institutions affiliated with the UK Reproducibility Network. The data provide a snapshot of open research practices that can be used to identify barriers, training needs, and areas that require greater investments.

Abstract Biological science produces “big data” in varied formats, which necessitates using computational tools to process, integrate, and analyse data. Researchers using computational biology tools range from those using computers for communication, to those writing analysis code. We examine differences in how researchers conceptualise the same data, which we call “subjective data models”. We interviewed 22 people with biological experience and varied levels of computational experience, and found that many had fluid subjective data models that changed depending on circumstance. Surprisingly, results did not cluster around participants’ computational experience levels. People did not consistently map entities from abstract data models to the real-world entities in files, and certain data identifier formats were easier to infer meaning from than others. Real-world implications: 1) software engineers should design interfaces for task performance, emulating popular user interfaces, rather than targeting professional backgrounds; 2) when insufficient context is provided, people may guess what data means, whether or not they are correct, emphasising the importance of contextual metadata to remove the need for erroneous guesswork.