Ask HN: What tools should I learn to make explorable explanations?

[MooMooMilkParty]

Primarily I'm interested in making things like Bret Victor and Nicky Case, but with an emphasis in my area of work. My background is in computational Earth sciences, and I mostly use Python (numpy, scipy, matplotlib, pandas, xarray, etc).

I've considered using Plotly Dash, but would love to be able to do these things with client side computation so I don't have to fuss around with hosting anything.

What I'm looking for:

- A workflow (notebook for prototyping, etc)

- Numerical libraries (numjs looks promising, but would love more advanced functionality)

- Visualization libraries (is anything in competition with D3?)

- A nicer way to write Javascript that interacts well with above (Coffeescript syntax seems nice, Typescript seems to have more hype?)

Is there any sense in trying to go all out and write my analysis code in Julia or Numba (via Python), and spit out the LLVM IR to feed to Emscripten and compile to JS?

[ChrisRackauckas]

>Is there any sense in trying to go all out and write my analysis code in Julia or Numba (via Python), and spit out the LLVM IR to feed to Emscripten and compile to JS?

Julia is working towards something like this. This GSoC project description is pretty clear about the current state of web assembly:

https://julialang.org/soc/projects/general.html#webassembly

[MooMooMilkParty]

I hope this capability gets to maturity. It consistently seems like anything I want in a language is in progress in Julia; new scientists in 15 years are going to be very spoiled if Julia becomes mainstream.