From 3DGeo sprint:
The function show_interactive claims to “visualize the dataset in JupyterLab”. Actually, a raster model with a given cell size is created, which is then shaded or which a slope map is generated from. The workflow should more exactly be Point cloud -> Rasterization (here we also have a couple of options, like min/max/q95/…) -> Shading/Slope map. Further, the Hillshade includes a TIN filling of the convex hull (I guess?), could this also be an extra step in the pipeline?