Voxel Plotting with Gnuplot 5.4

lwn.net

108 points·leephillips··8 comments

Nice, I've been looking for a way to visualize the evolution of a volatility surface over the course of a month, the trick of just outputting the files and greating a gif with ImageMagick seems like a nice, quick way to show it. Any idea how you could make something like this but interactive, say within a Jupyter notebook with a slider?

Really great work here, I need to renew my expired LWN subscription clearly. Thanks for posting!

ipyvolume and plotly can do some similar things.

You absolutely could do this. Here is another article I wrote for LWN with an interactive graph:

https://lwn.net/Articles/746386/

It uses matplotlib, but there is also a gnuplot kernel for jupyter. It would take some work to hook everything together, but I'm sure it's possible.

VTK is a great visualization tool that might suit your needs.

If you want to dig into VTK, try out ParaView (https://www.paraview.org/) or VisIt (https://wci.llnl.gov/simulation/computer-codes/visit), both open source and scriptable with Python.

VTK is a mature industry standard for 3D data serialization and visualization. That's much more heavy then mplot3d (matplotlib 3d) or gnuplot 3d.

3D viz that integrates well with Jupyter: https://yt-project.org/

Also MayaVi (https://docs.enthought.com/mayavi/mayavi/index.html), which is more Pythonic.

Paraview (https://www.paraview.org/), which is widely used and has a relatively good GUI, but not as good for Jupyter.

Ascent (https://ascent.readthedocs.io/en/latest/) for visualization of in-situ data.

You can use ParaView from Jupiter just fine ? You can even embed its views, pipelines, controls, make interactive notebooks, etc...

Voxels are the 3D equivalent of pixels

Reply on news.ycombinator.com