Why Reproducibility Matters? A Personal Experience

<ul> <li> <p>How many times have you got frustrated, because your code from last month refuses to run and your past self didn't bother to leave you any guidelines?</p> </li> <li> <p>Are you scared that one day a colleague will ask you the data/results of a paper, and you will have to excavate them out of a folder named 'allDataResultsVersion8bFinalizedRecheckedPleaseEnd'?</p> </li> <li> <p>Are you concerned that your contributions will not have much impact, because people may not be able to access them properly?</p> </li> </ul> <p>If yes, you are in the same shoes with me, when I started my PhD! In this talk, I present how I embraced reproducibility to avoid some of these common problems. I explain some best practices, simple tricks and habits I have learned throughout my doctoral research, which improved the quality of my research, made it reusable by others (and myself), and hence improved the accessibility and visibility of my work. I give practical examples in where I failed (and then gradually improved) on organizing, developing, versioning, documenting, licensing and publishing the research material (data, code, experimental setups and results etc.). I also introduce some of the available software tools and services, which would help you to achieve these goals. I hope that this talk will convince you to consider reproducibility as a major criteria in academic research, and give you a head start on achieving this objective.