The pair correlation function (in a sense, the distribution of interparticle distances around a probing particle) is useful in the case of monodisperse monomers (3D-particles, 2D-discs), when you are interested in short-range or long-range correlations in particle positions, regardless their own polydispersity. Additionaly, to characterize the aggeragtion state, you can use 2D cluster-size distribution in terms of number-frequencies  for clusters of a given size (particle number).

If you deal with a polydisperse ensemble, you have to characterize the particle size distribution itself, and, in addition, the 2D-spatial distribution of those polydiperse monomers. The first one will give information about the particle ensemble in terms of average size and dispersion. This distribution can include also the aggregate fraction. Again, the spatial distribution can be characterized in terms of pair-correlation function.

I remeber an excellent book explaing the pair-correlation functions: C.A. Croxton: Liquid state physics – a statistical mechanical introduction. (Wiley, New York, 1986)

About the cluster-size distribution- perhaps these reviews can help:

 Meakin P. Fractal aggregates // Adv. Colloid Interface Sci. 1988. V. 28. N. 4. P. 249-331

Models for colloidal aggregation // Ann. Rev. Phys. Chem. 1988. V. 39. P. 237-267.

I've also crossposted this question in Gwyddion-users maillist and got a lot of relevant links there:

https://sourceforge.net/p/gwyddion/mailman/message/35361613/

OK,

good luck!