In order to make further progress down the "3D visualization" road a number of interesting topics - which the scheme of this task unfortunately did not allow us to persue - could be investigated.
The visual impression of the structures is - as mentioned above - strongly dependent on the way one looks upon the data, but also on the completeness of the samples. These two effects can be hard if not impossible to distinguish and it is therefore of vital importance that the completeness of the samples used in a visualization is investigated and documented.
Before this is done any quantification of the distribution could easily give in the best case inaccurate and in the worst case misleading results.
Several ways of realizing such a quantification could be considered. One could perform galaxy counts within small subvolumes of the complete sample and this way map the density distribution of galaxies. This density could in turn be visualized by means of isosurfaces or strings tracing the density.
Analysis of this kind may be an hithero unknown source for finding constraints on cosmological models and are therefore an obvious goal for further developments.