In this case, point cloud data derived from in-the-field LIDAR scans of foliage was processed by a third party script to produce multiple cylinders approximating plant structures. Using Python to import the output of this script, Blender can display both the point cloud and cylinder-fits in the same scene, and permit the tweaking of the cylinders to best correct problems with the initial script output.
Here the fact that both Paraview and Blender support deep scripting using Python enabled us to save frame-by-frame scene data from Paraview in a form that could be imported into Blender. Paraview data was converted to a series of OBJ files which were in turn imported to Blender using a Python script that assembled them into a mesh sequence suitable for high quality rendering. In this case the aim was to provide a more impactful visual display of the results of a physical simulation than was possible using Paraview alone.