What about other sources like Google Scholar?

Also Dimensions or Web of Science. Constraining this to one source, that just happens to be owned by the sub-contractor, is an exceptionally bad practice.

Agreeing with the comments: WoS, Dimensions, and Google Scholar must be added.

Other than that, citation counts have nothing to do the the form of Open Access.

And OpenCitations is missing and must be added.

Number of citations is a very tricky measure because it is associated with a time window: it takes some time for papers to be cited, which is largely dependent on the field of research, and on the type of findings. For instance, some truely groundbreaking work may take years to be understood by the community and only becomes cited after a long period of time, while some more minor incremental work may get highly cited if it is published right when a topic is most popular. In other words, number of citation in a short period of time (e.g 2 years) is a poor metric of how useful a publication is.

Dealing with citations you can't stick only to Scopus. Web of Science and other databases show different numbers. And, again, you are referring only to journal articles and not to "publications" in general, and only to STEM disciplines

Using a single, proprietary database to measure citations is a wrong method. All other available databases or tools should be added. It has to be underlined that again monographs and most of the SSH journals are excluded from any citation database.

Questions is what kind of statements you intend to base on this variable. To what extent is this valuable for assessing the adoption, penetration or effect of open science?