search

SIESTA-eu / wp15

work package 15, use case 2
0 stars 2 forks source link

update README flowchart #8

Closed robertoostenveld closed 3 months ago

robertoostenveld commented 7 months ago

Yesterday @marcelzwiers made some additions to the general README following discussions in the wp15 online meeting. He came up with this draft flowchart:

  1. data owner -> sends data to product owner
  2. data user -> requests anonymous data from the product owner
  3. data user -> sends analysis pipeline to the product owner and requests pseudomized or anonymized output data. This output data is produced by the data user's analysis pipeline running on the source data that is pseudomized (to a necessary degree) by the product owner (using software developed in this WP)
  4. product owner -> requests review and permission from the data owner to send the pseudonymized or anonymized output data to the data user
  5. data user -> repeats step 3-4 until pipeline is finished
  6. ??

I suggest to refine this to

  1. data owner -> sends input data to the platform
  2. data user -> requests the product owner for access to the platform
  3. data user -> installs software and dependencies
  4. data user -> requests the product owner for pseudonomized data to be disclosed (using tools developed in this WP)
  5. product owner -> requests the data owner for a review and permission to disclose the pseudonomized data to the data user
  6. data owner -> grants permission
  7. data user -> interactively implements and tests analysis pipeline on pseudonomized data
  8. data user -> requests the product owner for the pipeline to be executed on the source data, output data is not yet disclosed
  9. product owner -> requests the data owner for a review and permission to disclose the output data to the data user
  10. data owner -> grants permission
  11. data user -> uses output data to answer research question and publishes research outcomes
vincent-legoll commented 6 months ago

@robertoostenveld Would it be useful to make the input data scrambling step explicitely in the flow chart ? probably between 4 & 5, or as being part of step 5.

BTW, I think the "scrambled" -> "pseudonymized" is a good move

robertoostenveld commented 6 months ago

The process of scrambling is indeed implicit in-between 4 and 5. I have been a bit sloppy in describing the various actions. I think that they all fall in either: requesting something, doing something that involves data+compute, reviewing and (dis)approving something.

I don't want to renumber the list right now, but will add it to step 5.

robertoostenveld commented 3 months ago

the list that describes the flowchart is still overall correct, but does not include the separation between single subject and group analyses, nor the resampling and noise calibration. However, it is interesting to document the procedure not only from the technical side (i.e., all steps that are done and seen by the SIESTA developers), but also from the researchers side (all steps that the researcher takes) and separately from the data rights holder perspective.