L’accès public (comme l’option sur Github) est l’un des concepts clés de la science ouverte, qui vise à distribuer les results of the research (e.g., academic scientific papers) to the public to accelerate research, without fees or other barriers to access. It has been adopted by a growing number of researchers and agencies across a variety of fields. At the same time, software developed during the scientific research process is often released as open source software (OSS) to promote reproducible research and increase the likelihood of greater impact. OSS can be copied and distributed at virtually no cost, potentially opening the door to unprecedented levels of sharing and collaborative innovation.
The reproducibility challenges and artifact trails that have been hosted at international conferences have strengthened the links between software and scientific papers. It can be argued that a paper is more likely to have real impact if its code is accessible; readers are able to examine the code, run the code, and evaluate the contributions by considering all the details of the work (which are often omitted from the paper).
Python is the primary language used in data science, and there are many different IDEs you can use to code in Python, but to create a structured report with clean code and annotations, Jupyter Notebooks are the way to go. Here are some tips to make your notebook more accessible:
Use Markdown tags : si votre code est long ou complexe, il serait très utile d’avoir des titres, des sous-titres et des descriptions Markdown pour les morceaux de code sur lesquels vous avez travaillé. En procédant ainsi, une personne non technique ou qui n’est pas aussi experte que vous en codage peut comprendre pourquoi de telles mesures sont prises et ce que font certaines cellules.