Since the simulations consist of a limited number of point particles a number of steps have to be performed in order to create a continous distribution.

    The first step is to bin the data. An iterative method similiar to the TREECODE is used to create bins with close to the same number of particles in each bin.  As an example the primary galaxy would be binned as seen below.

    The much higher density in the center allows for smaller bins being used there resulting in a higher spacial resolution.

    After binning any number of quantities can be calculated in each bin. For these simulations the density and average velocity was calculated as well the velocity distribution and the third and fourth Gauss-Hermite terms. These latter quantities measure the deviations from a Gaussian distribution. Calculating them requires a seperate two step fitting process: a non-parametric penalized likelihood fit the line-of-sight velocities followed by a chi minimizing fit to the Gauss-Hermite terms.

    In order to reduce the noise across the projected plane a spline is be applied to produce the final 2D map.