Establishing a reproducible and sustainable analysis workflow

Getting started on any project is often the hardest thing - and when it comes to starting your career in research, just figuring out where and how to start can seem like an insurmountable challenge. This is particularly true at this moment - when there are so many programming languages, programmes, and systems that are freely available to neuroimaging researchers, and even more guides, tutorials, and courses on how to use them. This chapter is intended to set you off on the right foot as you get stuck into the task of learning to work with large neuroimaging data. We will cover a number of processes, systems, and practices that you should adopt to help ensure that your work is efficient, your processing steps traceable and repeatable, your analyses and findings reproducible, and your data and processing scripts amenable to sharing and open science. While this chapter is aimed at those getting started, it will also be of use to established researchers who want to streamline their processes and maximise robustness and reproducibility of their neuroimaging analyses. Finally, this chapter is also intended to help make neuroimaging work practices and processes more environmentally sustainable by reducing demands on computational resources through better planning, efficiency, and awareness of resource use.