On Manovich's account of data visualization's representation of the anti-sublime, it did seem fitting that data visualization should both renounce and return to prior practices of representing phenomenon.
I agree that it does perform the acrobatics of returning to a tangible representation of a thing, albeit now turning our attention to things previously perceived as transcendental of human explication.
From this view, I see data visualization not as anti-sublime, although the use of the term is eyecatching, but as uber-sublime.
Data visualization seems to be a vehicle from which you can make patterns, realities, truths that seemed humanly ungraspable ironically common-sense, within our field of vision.
Where "übermenschlich means super-human, in the sense of beyond human strength or out of proportion to humanity" (Wikipedia) -- as I understand it, not "superhuman" as you would correlate with a Superman or Batman, but as in an extended or augmented human faculty -- data visualization could challenge the idea that there are phenomenon that transcend human perception, or that are sublime. The term referenced carries with it some notions of ideas that could go terribly awry from an evolutionary perspective.
From the media-culture advancement perspective, however, I do see visualization creating, designing novel ways human beings map information cognitively. It was very interesting when Manovich brought up visualization embracing meta-media in its performances, thereby bringing in other dimensions of information into the physical.
This question of how multi-dimensional visualization can impact our current arrangement with information interpretation and dissemination that I find most intriguing - more so than Manovich's other probes such as who and how/why are visualization strategies employed.