Pretty much like Jupyter notebooks. It's a great way to present research, especially statistical stuff. It's pretty much impossible to verify any statement about a large dataset without having it to hand and ready to calculate. Even having the data printed in an appendix is going to be discouraging verification.
Plus the stories you can tell with this sort of thing are much more vivid. If someone has a question about some part of the data, they can ask the computer instead of the author. For a lot of queries.
The article even has an image of the interface from 1988.
> And with the release of Mathematica 1.0 in 1988 came another critical element: the invention of Wolfram Notebooks. Notebooks arrived in a form at least superficially very similar to the way they are today (and already in many ways more sophisticated than the imitations that started appearing 25+ years later!): collections of cells arranged into groups, and capable of containing text, executable code, graphics, etc.
Wolfram really is a terrible liar (also very, very smart). He knows damn well that Maple had notebooks in the mid-1990s, because it was one of Mathematica's main commercial rivals, so "started appearing 25+ years later" is nothing other than a lie.
I considered that interpretation before writing and I found it impossible to read that way. For a start, "imitations" is plural, so no, it doesn't just refer to Jupyter. And "started" certainly implies that other imitations hadn't appeared before that.
While I'm at it: he deserves an uncharitable reading, because he's having a dig at Jupyter (which may be an imitation of Mathematica notebooks) and neglecting to mention any influence of previous work on Mathematica notebooks, such as Knuth's literate programming.
Plus the stories you can tell with this sort of thing are much more vivid. If someone has a question about some part of the data, they can ask the computer instead of the author. For a lot of queries.