Plagiarism is one thing, and I support your view on this situation as paraphrasing data. But a more important reason to add a citation is to let the reader know where the data is coming from (and how the data was collected). This is immensily important for the reader to understand and analyse your presented data correctly. In my mind, we should always write with open cards, i.e. so that our work is repeatable and deductions are easy to follow deductions.
Should I just simply provide an in-text citation in the caption and not mention that the figures won't be found in the cited source?
I would recommend a sentence in the figure text of something like: "The data in this figure are derived from raw data found in [source]." Preferably add an equation reference to the equation that the plotted data is a result of: "The data in this figure are derived from raw data found in [source] by using (equation reference)." or ..."by applying (equation reference) to it."
In my mind this kind of sentence will make it clear where the raw data is coming from, and that the figure you present is your work and therefore not found in the source.
Should I, perhaps, forgo citation of the data's source in the figure and only cite the data's source when discussing the figures in the actual text?
The reference to the raw data should be in the figure text. A common way to read a report/article is to read summary and conclusion and scroll through to read highlights and look at figures. Therefore the figure texts should explain their figure in a quite independent way, so that the quick read gives a good understanding of what the figure presents without searching through the main text. References to how the data was collected and computed, is part of the foundation for understanding the figure.