Matplotlib 3D and AVP?

[chars1ub0w]

I write Python code to generate 3D graph objects in Matplotlib. Just a simple example of what you can rotate around on a regular 2D screen.
(Seems I cannot attach the movie file here.)

I'm wondering if there is a browser plug-in or something that permits one not to just rotate the 3D graph, but to actually move through it. And if so, would it make a nice demo on the AVP? Unfortunately, I don't know of anyone who has an AVP.

 

[Slartibart]

you can easily use Matplotlib‘s interactive navigation here - or do I misunderstand you?

 

[chars1ub0w]

you can easily use Matplotlib‘s interactive navigation here

Yes, of course, being a user, I know that. I was hoping there was something special, a AVP viewer for 3D objects that permitted one to walk through in an immersive way. Especially, if the graph is quite complex, telling them how to walk it a certain way may show graph relationships that are not obvious from a single glance. Then it'd be a nice "toy" for garnering interest in a field, something that college and high-school students might line up to experience. That kind of thing.

 

[Slartibart]

you can export your matplotlib 3D plot as STL and convert to USDZ, which then will be „accessible“ in e.g. Quickview on all Apple incl. the VP. You might do some triangulation on your 3D plot to improve visuals - but you are using python already, so that seems trivial.

i am curious - what do you expect to better accessible in such an approach, then a different projection to, or a transformation of the data, to show relevant data structures in 2D? What type of data, which field are you talking about?
I can see how something like this can appear impressive for none-professionals, but in case of complex visualisations, it might be much more to the point, to reduce complexity instead of walking - even in some form guided - into it. 😃

But in general, what you want to do is technically solved - plot, export, and then, as a start, open in Apple‘s Reality Composer and apply some transform animation, interactivity, location awareness. labels…

 

[chars1ub0w]

you can export your matplotlib 3D plot as STL and convert to USDZ, which then will be „accessible“ in e.g. Quickview on all Apple

Thanks for the pointer to USDZ.

i am curious - what do you expect to better accessible in such an approach, then a different projection to, or a transformation of the data, to show relevant data structures in 2D? What type of data, which field are you talking about?

Before I go writing more code, I was hoping for something that'd be more immersive than simply rotating a graph in a window. For example, if one could move through the graph object, turn one's head in some direction and see what's to the side, above and below.

As for what insights etc., the eventual idea would be to de-complexify a graph object that might contains many thousands of nodes and types of edges by turning your head. Idea being that if you turn to certain angles, certain types of edges would appear or disappear. And if you find the right angle, you see a path through to the node we're looking to find. That would make it interesting to college students and high schoolers in terms of a wow factor and also for them to learn about the different kinds of links between nodes. Useful for demo (of solved cases).