netclop

NETwork CLustering OPerations (for geophysical fluid transport).

netclop is a command-line interface for constructing network models of geophysical fluid transport and performing associated clustering operations (e.g., community detection and significance clustering).

Features

• Binning of Lagrangian particle simulations using H3
• Network construction of LPT connectivity
• Community detection using Infomap
• Network resampling and recursive significance clustering
• Node centrality calculation
• Spatially-embedded network visualization

About

netclop was created as a CLI to facilitate network-theoretic analysis of marine connectivity in support of larval ecology. It functions as a library to computations on network ensembles. Developed at the Department of Engineering Mathematics and Internetworking, Dalhousie University by Karsten N. Economou.

Papers

• 2024 - Characterizing variability in complex network community structure with a recursive significance clustering scheme (Karsten N. Economou, Cassie R. Norman, Wendy C. Gentleman)

Usage

CLI

netclop accepts Lagrangian particle tracking (LPT) simulations decomposed into initial and final positions in as .csv structured as

initial_latitude,initial_longitude,final_latitude,final_longitude

as an input. Recursive significance clustering is run on all provided filepaths of LPT position files and stores all produced content in the specified output directory

netclop rsc [OPTIONS] [PATHS] -o [DIRECTORY]

If one LPT position file is given, it will be bootstrapped; otherwise, each LPT position file is treated as an observation.

Significance clustering

Significance clustering can be run on a networkx.Graph object directly, which will partition and bootstrap

from netclop import NetworkEnsemble
ne = NetworkEnsemble(net, **ne_config)
ne.sigclu(**sc_config)
cores = ne.cores

or on an ensemble of partitions

from netclop import SigClu
sc = SigClu(partitions, **sc_config)
sc.run()
cores = sc.cores