In the limited 2 hours which was available we were able run the script and check some analysis was computationally reproducible.
I am a statistician so am familiar with methods and R
R studio, R
Encoding problems, changing between I assume Mac and widows lead to problems. The problems occurred because of different versions of dealing with symbols, these problems could be avoided by using clean text in Janitor, to clean the data names. It is a good data practice especially with final csv files for analysis to never have spaces or symbols in the name, for example you have Phase (c=Intervention), my computer did not read it in the same way as yours. Also some of the name of variables are sentences, it would be better if they had meaningful short names
Spelling mistake in installing GPArotation.
This is not fully reproduceable as the top of the file need to be replaced to map the drive. This is the older way to do this, while correct it breaks reproducibility, a much better way to achieve this is to use projects within R.
I would suggest not including the install packages as this can can cause issues for other people's installation. R tells asks you to install anything that is missing so it is not a big issue anymore. Only load the packages once during a session, Remove any redundant code.
Note the dates in your files are not in a good format. please concider using yyyy-mm-dd
The syntax was clear and was easy to follow with comments
I can see the authors have tried hard to make this work reproduceable. It would not take much more effort to make this fully reproducible using RMarkdown.
Could be improved by using projects and hithub and R markdown so it then could be just cloned and run.
Overall, there was a clear structure to find the materials.
the use of R markdown would make it easier to understand which analysis went with what part of the paper.
Use good data practices, such as name and date conventions as described above