Recommendation for Adopting Open Science

This document is under development. References will be added soon.

Data Science Without Borders (DSWB) is committed to applying principles of open science for facilitating collaboration, transparency, reproducibility, and accessibility in health research across Africa. This document outlines how DSWB will implement these principles, maximising the impact of data science in health research and ensuring that knowledge through DSWB is shared equitably and openly, including the requirements of our funder, Wellcome Trust.

These recommendations apply to all research activities and outputs within the DSWB framework, including:

• Research projects
• Data collection and analysis
• Software and tool development
• Publications and presentations
• Capacity building and training activities

For introduction and relevant details about open science practices, please refer to the Open Research Chapters in The Turing Way and browse resources shared on Zenodo: https://zenodo.org/communities/dswb-open-science/.

Important Considerations¶

• Open Access Publication, Open Data and Open Source Software: All research outputs, including publications, data, and code, should be freely available and accessible to everyone.
• Open Sharing: Research processes, methodologies, and results should be transparent and well-documented. All research outputs should be shared openly to enable validation, replication, and further analysis.
• Reproducibility: Research findings should be reproducible, with clear documentation of data and code, such as via research articles, executable notebooks and GitHub repositories.
• Open Collaboration: Foster collaboration and knowledge exchange among all DSWB Consortium Delegates and members from their networks, including researchers, practitioners, and communities.
• Equity and Inclusion: Ensure that the benefits of open science are shared equitably, and that all stakeholders, regardless of their location or resources, can participate in and benefit from DSWB activities.
• Funder Compliance: Adhere to the open access and data sharing policies of the funding organisation, Wellcome Trust.

Open Access to Publications¶

• DSWB requires that all research publications resulting from Wellcome Trust-funded research are made open access in accordance with the Wellcome Trust’s open access policy.
• DSWB encourages the publication of research findings in open-access journals or platforms indexed in the Directory of Open Access Journals (DOAJ), as per Wellcome Trust guidelines.
• In cases where open-access publication is not feasible, DSWB will ensure that publications are made available through other means, such as institutional repositories or self-archiving, in compliance with funder policies.
• DSWB supports the use of Creative Commons licenses (CC BY or CC BY-SA) for publications to facilitate reuse and adaptation, with a preference for CC BY as per Wellcome Trust policy.
• Authors will ensure they retain sufficient rights to comply with open access requirements.

Open Data¶

• DSWB promotes the open sharing of research data, while adhering to ethical, legal, and privacy considerations, and in line with the Wellcome Trust’s data, software, and materials management and sharing policy.
• Data should be Findable, Accessible, Interoperable, and Reusable (FAIR).
• Any restrictions on data sharing should be clearly documented and justified.
• DSWB has established guidelines for data management and sharing, including the use of data repositories and metadata standards. Please refer to the DSWB’s Data Sharing Agreement for details.

Open Source Software and Code¶

• DSWB is committed to the open sharing of software and AI models developed through different initiatives led and supported by the consortium partners.
• Software and code developed within DSWB projects should be released under permissive open-source licenses (e.g., MIT, Apache 2.0) to facilitate reuse, modification, and distribution.
• DSWB will promote the sharing of trained AI models, including model architectures, weights, and training code, to enable reproducibility and further research.
• DSWB will encourage the use of version control systems (e.g., Git) and code repositories (e.g., GitHub) to facilitate collaboration and transparency in software development.
• Documentation and metadata will be provided alongside software and AI models to ensure their proper use and understanding.

Transparency and Reproducibility¶

• DSWB is committed to transparency in the research process. Researchers should document their methodologies, data sources, and analytical procedures in a clear and comprehensive manner.
• DSWB promotes the use of tools and practices that enhance the reproducibility of research findings, such as:
  • Preregistration of study protocols
  • Use of reproducible research environments (e.g., Jupyter Notebooks, R Markdown)
  • Documentation of data provenance
• DSWB will encourage the publication of code and data alongside research publications to facilitate reproducibility.

Collaboration and Knowledge Exchange¶

• DSWB seeks to foster a collaborative environment that encourages the sharing of knowledge, expertise, and resources.
• DSWB will organise meetings, workshops, conferences, and other events to facilitate networking and collaboration among researchers, practitioners, and communities.
• DSWB will promote the use of online platforms and tools for collaboration and communication.

Capacity Building and Training¶

• DSWB is committed to building capacity in data science, with a focus on open science practices.
• DSWB will provide training and educational resources on open science principles, data management, and reproducible research.
• DSWB will support the development of open educational resources (OER) to promote open science education.

Ethical Considerations¶

• DSWB recognises the importance of ethical considerations in research, particularly when working with sensitive data or vulnerable populations.
• All research activities conducted within the DSWB framework must adhere to the highest ethical standards and comply with relevant regulations and guidelines.
• DSWB recommends that each Pathfinder Institution follow the ethics review process of their respective organisation to ensure that all research proposals are evaluated for their ethical implications.
• Informed consent should be obtained for any data collection involving human subjects.
• DSWB will ensure that data is collected, stored, and used in a manner that protects the privacy and confidentiality of individuals and communities.

Open Science and Capacity Building Working Group¶

• The DSWB’s Open Science and Capacity Building working group will drive the implementation of open science and ensure compliance with the Data Sharing Agreement.
• In DSWB, technical partners, specifically the OSPO Now team, will provide guidance and support to researchers and staff on how to implement this policy in their work.
• DSWB will provide metrics and indicators to monitor the progress and adoption of open science practices to assess the impact.

Dissemination and Advocacy¶

• DSWB members, supported by technical partners like OSPO Now, will actively promote the principles of open science and advocate for their adoption within the broader global health and development community.
• DSWB members will disseminate their projects and open science guidelines applied in their work through their website, publications, and other communication channels.
• DSWB will engage with policymakers, funders, and local stakeholders to promote open science and data sharing.