Interactive 3D visualization of data is a crucial part of the analysis toolbox for many subfields of computational science and engineering, ranging from fluid dynamics to biology, mathematics, cosmology, and more. With the widespread availability of WebGL in browsers, it is now possible to include GPU accelerated 3D graphics directly in Jupyter Notebooks (www.jupyter.org). As part of the OpenDreamKit project (www.opendreamkit.org), we are working to improve the state of 3D visualization in Notebooks.
In the last couple of years a number of visualization projects have started to explore this area, including established projects such as VTK, Paraview and MayaVi, as well as newer, smaller projects such as ipyvolume, pythreejs, k3d-jupyter, SciviJS, and unray, the latter three being initiatives from the OpenDreamKit project.
In this talk we will present the current state of 3D visualization in Jupyter Notebooks and our contributions, including our work to lessen fragmentation and double-work in the field.