Thank you for your comment. I have described one possibility for pseudonymised data, and described how this could be achieved with GRAKN.AI (and any graph database for that matter) in this second article.

For me the chalenge is re-identification, as I tried to show in the previous article, it is really easy to re-identify someone, by crossing data with other sources, and honestly, I do not see how you can prevent it. I would be very happy to have someone propose a solution.