If any part of your work is open it is actually more difficult to be scooped as they are often time stamped. Timestamping is not sure-fire. A researcher could scoop you and publish in journal X. Journal X may not care about your timestamp and refuse to retract. If you then try to publish in journal Y, journal Y may say 'this is already published in journal X'. Suggested response to this: Journal or university lawyers can help you get scooping articles retracted if you can prove it, and time stamps will help with that.
You could still share meta-data/information about what data you have. Would still be useful to researchers to avoid double-work, and they could contact you to ask for specific data, which would make reuse at least possible in principle.
You could share data after you have published. If you share data bit by bit over multiple papers, resources like DANS can help you link data across multiple projects together.