I am not sure but there seems to be a slight tendency towards visual programming these days [2-4].
It might be relevant to @evancz’s explorations as mentioned in [1].
My hope was to focus specifically on scientists and researchers, so it’d be a package that makes it easy to quickly see a bunch of regressions or the impact of changing a constant in a formula.
A lot of the big tasks there are in loading data formats in a nice way.
[1]
Have you considered the history of orange [5]? Some relevant excerpts from the podcast:
08:30 visual interaction is more powerful […]
09:13 in 1993 I had a class in high performance computing with a tool named data explorer, something similar tool by silicon graphics and it used visual programming. You could design the workflow how you would then analyse the mission[?]. […] you could change the parameters. This was fantastic. In 1993 this tool existed. That doesn’t exist any more. And then in 2003 we said that’s the proper way to address it.
09:50 We need a visual programming framework, where you construct data analysis workflows very easily. Just by putting the blocks together. We called them widgets. And we need heavy visualisations. That means everything needs to be visualised. Not only data but the models. And we need interaction. So in any kind of visualisation you can touch any element and then you can find out what particular data is actually associated with that.
Why not something like ellie-app specifically for data analysis + interactivity?
Sort of Jupyter lab experience with visual programming where each lab sheet includes not only the workflow but also the data itself.
[1] https://twitter.com/czaplic/status/1254901118108393472
[2] https://www.luna-lang.org/
[3] https://orange.biolab.si
[4] https://darklang.com/
[5] The Python Podcast.__init__: Orange: Visual Data Mining Toolkit with Janez Demšar and Blaž Zupan