Work in collaboration with Grzegorz Siudem and Przemysław Nowak from @WUT_edu, and @chakresh_iitgn @luyibov at @research_lpi

The method, using a 2-parameter DGBD distribution and a binning method is particularly useful for concave distributions and large datasets (N>100). In particular, it avoids the fit being dominated by low-rank values in the tail.

New paper from the team just published in Physica A! We propose a novel fitting method for rank-size distributions beyond Zipf's law, and apply it to the case of commit distributions of open-source repositories in Github. sciencedirect.com/science/ar…