netsegpol 0.1.0

Structural polarization and segregation measures for networks.

NetSegPol

This repository provides a lightweight, object-oriented Python toolkit for computing structural polarization and segregation measures in networks. It implements a wide range of classical and modern network mixing, segregation, and polarization indices used in sociology, computational social science, and network science.

How to use?

You should not use descriptive measures without a null-model.

from netsegpol import Measurer

edges = [
    # Group 0 
    (0, 1), (0, 2), (1, 2), (1, 3),
    (2, 3), (2, 4), (3, 4), (0, 4),

    # Group 1 
    (5, 6), (5, 7), (6, 7), (6, 8),
    (7, 8), (7, 9), (8, 9), (5, 9),

    # Sparse inter-group edges
    (2, 6), (3, 7), (1, 8), (4, 9)
]

membership = {
    0: 0, 1: 0, 2: 0, 3: 0, 4: 0,
    5: 1, 6: 1, 7: 1, 8: 1, 9: 1
}

measurer = Measurer(edges, 
                    membership,
                    directed = False,
                    safe_create = True)

print(measurer.segregation_matrix_index()) # or the function that you want to use

What should I be careful of?

You should be careful if you are using an external package like igraph and getting the edge list directly since indexing methods might differ and your edgelist and membership dictionary might not match.

Will there be new measures?

This is still a project in its infancy, hence a few of the measures will be developed later on, such as Ergodic Markov Chain mixing time or Spectral Segregation etc.