Statistical analysis tools for particle quantities

[Hendrik1704]

I envision a class, where one can put in a particle data list and this can then compute some useful quantities like mean multiplicity or mean transverse momentum of the provided data set.
If the list contains multiple events, one could even return the standard deviation or higher moments of the distribution.

A mean ncoll could be also interesting for transport computations, as this is related to the Knudsen number.

@nilssass, @NGoetz, maybe you have further thoughts on this or even ideas for other quantities to include there.

[NGoetz]

Generally what we offer is already quite close to this, as one can easily extract this from the architecture we provide. It gets more interesting though, if we add higher moments of distribution and stuff out of the box, as this increases usability if we design the class nicely.
For my personal plans with SPARKX though, I would prefer focusing on #15 first, and then implementing such a class in the next step.

[Hendrik1704]

Sure, #15 has the highest priority for the next release(s). I just wanted to fix my thought somewhere, such that I can still remember this at some point.
I would also say that the implementation of higher moments would be interesting and it is something other codes don't have.

[NGoetz]

I think it is a good point now to start looking into this too. Using SPARKX I realized that though such things are easily implemented with SPARKX on usage, the downside is that one has to test all of this then 😆
However, in such an implementation it is essential to allow for multiple quantities to be computed in the same loop, in order to improve performance, and to avoid looping through large data files multiple times.