Data classification is important. Open Access implies access to everyone, without cost; thus, if there are any data that need to be kept private for privacy or business confidentiality reasons, they cannot be open access. In terms of scientific publications, this is actually the same in the subscription-based (closed access) model; all data must be sufficiently anonymized before publication.

Data classification is important. Open Access implies access to everyone, without cost; thus, if there are any data that need to be kept private for privacy or business confidentiality reasons, they cannot be open access. In terms of scientific publications, this is actually the same in the subscription-based (closed access) model; all data must be sufficiently anonymized before publication.

On the other hand, one component of "secure" is integrity, i.e., assurance that the data (or publication) have not been modified in an unauthorized manner. This can be accomplished by a digital signature on the data (or publication), either by the data owner or publisher. This is completely compatible with the Open Access model.

Once the data is open(in the sense available through any open license) it can't be private.

There are methodologies to release data that is private to open (research and analysis, etc.). The idea is that the data get sanitation: The Personally identified fields are eliminated (or encrypted) and the rest are chosen in a way that maintains privacy of individual records via privacy methodology. One such methodology is Differential Privacy, see https://en.wikipedia.org/wiki/Differential_privacy

In general data privacy is an important recent issue, to start see: https://en.wikipedia.org/wiki/Information_privacy