Hi, I’m working on a package for manipulating EEG (electroencephalogram) data. So far I’m working on data that is already pre processed by other software (i.e., fieldtrip, Brainvision), and the package implements dplyr-like functions for the EEG data (which is a list of several data frames, and because of size issues cannot be in long format), and in addition some EEG tailored function (re reference electrodes, baselining, etc). At some point I’ll add the pre processing functions to the package as well, so that the EEG pipeline can be started from scratch.
My question is first of all, is something like this in the scope of Ropensci? Can I submit a package is not in it’s “definite” form? Or should I wait until I have everything implemented? (I would really like to get feedback sooner rather than later, and that’s why I thought about a relatively early submission).
Seems like a really cool package idea and I’d guess that it’s in scope for rOpenSci.
Worth noting that a quick Google search pointed me to multiple R packages related to EEG analysis (though none of them seem to be in the rOpenSci ecosystem yet). This isn’t necessarily a bad thing, especially if those packages don’t include the functionality you’re after, but worth keeping in mind regardless.
I’d definitely suggest you start by reading through the documentation on contributing. Someone more experienced with onboarding may have a different opinion but I think you’re better off trying to get all the ‘basic’ functionality you can into the package before submission.
Thanks for the answer. Yes I’m aware of the other 3 packages (eegkit, eegUtils, erpr),but none was doing what I wanted
My project lives now in bitbucket, will I need to migrate it to github? I’ve already asked about continuous integrations with R and bitbucket (continuous integration with bitbucket), which doesn’t seem to be entirely straightforward. But reading through ropensci documentation, it seems that it only fully supports github.
Yeah you’re going to want to migrate to GitHub. As far as I can tell, the whole onboarding process is GitHub-dependent (open to being proved wrong though ).
Hi @bnicenboim! Yes, this seems likely to be in-scope for us. Packages designed to support data manipulation of scientific formats and data types such as EEG data make sense. See the “Package Overlap” section of our developer guide in regards to how we think of dealing with multiple packages covering similar areas.
We are pretty GitHub-centric, but we have had a few packages come through that are primarily developed on GitLab. I believe it is pretty easy to set up a GitHub-Gitlab mirror. Upon acceptance the package needs to move to our GitHub organization, but you can continue to administer and work on it via GitLab that way.
The appropriate stage for onboarding is always a challenge. It’s helpful to get early feedback, but we require sufficient feature completeness, testing, and documentation so that reviewers can give good feedback and sign-off on the package, even if there is a roadmap for more development in the future. If you still have core parts of the package whose design you are working in, it is probably not a good point to submit. However, you can still solicit feedback from our community on this! A good way to do so is post a link to your package on this forum along with some specific questions you’d like people to respond to.