And there's obviously been an increase in technical publishing of software matters: before we just had books, now he have books, plus blog posts, plus online white papers, plus videos, etc.
I don't see how an increase in the publishing of scholarly papers implies an expected increase in books, specifically.
Books and academic publishing are related, if not identical, activities. I'd expect there'd be some correlation. And the data for academic publishing happened to be handy.
I suppose its probably possible to extract data from other sources -- ibiblio, perhaps -- for books, but that would involve work.
There's also the matter that the xkcd infographic shows long-term trends. It's entirely possible that there've been changes in recent years. Though that would also be pretty remarkable.
The huge increase in online access to content and information has a bigger effect though. I was purchasing a large number of books through the mid 2000s, far fewer since. The fact that numerous physical bookstores I could head to and shelf-browse have closed doesn't much help matters (I hate buying anything online, books included, especially via Amazon).
I don't see how an increase in the publishing of scholarly papers implies an expected increase in books, specifically.