Thanks. Regarding your questions
Q.1. Yes but in 3D it is a sphere, or more precisely, a paraboloid approximation of a part of a sphere owing to the large radius of curvature compared to the contact area. The analytical Hertz solution of the tutorial corresponds to the sphere case. What I was saying is that in 2D (non axisymmetric) you cannot hope to model the contact of a sphere on a half-space, it will necessarily be that of a cylinder (or parabolic cylinder with a similar approximation) and that you need to compare against the Hertz solution of a cylinder against a half-space.
Q.2 I think I used in Paraview something like Sources > Geometrical Shapes > Sphere and clipped some parts for better visualization
Q3. Yes you can try combining the two approaches