7

I work and teach in a field where public data is scarce (anatomical MRI processing). I planned to organize a few practical works for students this semester. I had a pretty hard time gathering publicly available data from different software suites/open projects to organize a consistent and motivating practical work. Now, my question is the following:

Given that all these data are available to download (examples of data: fsl.fmrib.ox.ac.uk/fslcourse (Section Data Files)) on the author/lab website, do I "have the right" to create "my own data set" from parts of this data and under what conditions?

The reason that pushes me to do so is that some of the data comes from different heavy archives with hundreds of unnecessary things, with heterogeneous naming from one source to another, and I would like to "repackage" only the needed data in a consistent and comprehensive way and make this repackaged archive available somewhere for my students.

philshem
  • 17,647
  • 7
  • 68
  • 170
beuhbbb
  • 129
  • 3
  • 1
    Duplicate of https://law.stackexchange.com/questions/11359/can-you-copyright-data –  Sep 16 '20 at 09:00
  • @AnonymousPhysicist Thanks for the link. Does it suggest that I can do what I want with available raw data ? Is copyright the actual think I need to consider ? –  Sep 16 '20 at 09:25
  • Maybe medical privacy? In any case, this question does not seem to be about academia to me. –  Sep 16 '20 at 09:47
  • This depends a lot on what you mean by "make this repackaged archive available somewhere". You can't republish the copyrighted work of others, certainly, even if you reformat it. But you can usually use it for educational purposes. –  Sep 16 '20 at 11:07
  • @Buffy thanks. It seems to me it is contradictory with law.stackexchange.com/questions/11359/can-you-copyright-data. Does it ? In any case, I guess this is indeed an important point and maybe the best for me would be to provide tempory links to the data to my students (so that I am not republishing the data but simply providing modified version of already existing data, but the difference seems thin to me ...). Nonetheless I was quite enthousiatic to put the practical guideline as well as the data online so that anyone can do it, so more about this aspect would be appreciate :) –  Sep 16 '20 at 12:12
  • I am also in MRI processing so I am familiar, but your jump from the first para to second is confusing. You seem to be discussing a dataset you hadn't introduced. Can you clarify –  Sep 16 '20 at 15:52
  • Did you 'sign' a Terms of Service agreement when you downloaded the data? Many datasets have requirements that are stated when you download the data. Of course the chances that you'll be 'caught' violating the TOS are rather low. – user11599 Sep 21 '20 at 07:59
  • @user11599 Thanks for your input. No. For example some data are from here : https://fsl.fmrib.ox.ac.uk/fslcourse/ (Section Data Files). I just downloaded the zip files... and it was similar for the data I get. Do you think it adds some valuable information to my question ? –  Sep 21 '20 at 15:28
  • I suspect the Oxford data would fall under their software license. If so, it allows non-commercial use, but disallows commercial use. What would be your case? Do you have any "financial return" from compiling the datasets together? Following this approach you would have to check the license of any images you include. – mapto Oct 08 '20 at 14:12
  • @mapto. Thanks for your input. No I have no financial return (I guess, cause I am paid for giving the courses but this would strange to consider this is the case). So, if all data sets I used allow non-commercial use, in which way do you think I have to document the origin of the different data and licenses ? – beuhbbb Oct 09 '20 at 16:08

1 Answers1

0

This is fine. The only exceptions would be for identifiable data which would be covered by responsible conduct of research. You should take that training if you have not already in order to make sure you do not run afoul of the institutional IRB.