Planning for Project Design#

Thinking about reproducibility from the start of your project is the best way to save time and get the most out of the available tools. Creating connections between data, code, methodology, as well as diverse collaborators, can seem like a daunting task if it has not been planned from the beginning (see this illustration of Research Cycle). By documenting and sharing project workflows and processes for research and researchers we can ensure sustainability and reusability of research for both the developers and future users.

A human-centred approach in the context of the research project can lead to a better development process, maintenance, and future extension of our work. Furthermore, it will help improve the quality of future project design as we can learn lessons from what worked and what can be improved.

Setting Expectations Explicitly#

Project design practices help provide guidance and set clear expectations by explicitly communicating norms and empowering project members at different levels of hierarchy to collaborate equitably. A project design for reproducible research encompasses a variety of aspects, starting from defining the purpose, main research questions, expected users or target audience, available resources and skills required in the project. It also requires researchers to explore the possible outcomes, plans to address expected challenges or risks, ensure diversity of stakeholders and reduce possible barriers to participation.

Project design practices help all stakeholders to be certain about their roles and responsibilities, skill requirements, environment and research setup they want to create for their collaborators, values they want to promote and how they can achieve their goals collaboratively.

Getting Started#

We can begin with the project design process by identifying different parts of our research such as main research questions, methods and materials, code and data requirement, workflow, communication channels, ways of working, collaborative practices and so on. This process allows us to be intentional from the start to ensure that our research is reproducible, well communicated and inclusive of all stakeholders where decisions are collaboratively made. We can explore and select the right tools and methods for reproducibility in our research and promote good practices such as documentation, version control, peer-review processes, testing, workflow, archiving and data management plan can be promoted from the beginning. Finally, we can plan for publishing and sharing of research components before, during and after the project.

Organising Files and Documents#


Top three ‘selfish’ reasons to use project design practices

  1. Saves time: once the project is designed and all the connections between different parts of research can be organized with little effort.

  2. Makes your research openly available: having your research open from the start can help others working in similar subjects or starting research.

  3. Get people interested: you can get people to help you from the start since your project is documented and is easy to share.

Create a shared repository to allow easy access to information and different documents related to your project. A project repository can be openly available if you are developing an open source project, or can be shared only with your collaborators. In the chapter Creating Project Repositories, you can learn about how to set up a repository with key documents like a landing page, contribution guidelines and communication pathways.

Start by documenting your vision, mission, goals and roadmap in a README file (see the Landing Page - README File chapter), followed by a description of resources and norms for collaboration and communication. In the next chapter, we discuss how project design enables good communication and collaboration in research. In the subsequent chapters, we describe different methods and tools for ensuring research reproducibility.

Chapter Tags: This chapter is curated for the Turing Data Study Group (turing-dsg).