Identifying rich clubs in spatiotemporal interaction networks

Jacob Kruse a , Song Gao a,* , Yuhan Ji a , Keith Levin b , Qunying Huang a , Kenneth R. Mayer c · 2025

Read Identifying rich clubs in spatiotemporal interaction networks by Jacob Kruse a , Song Gao a,* , Yuhan Ji a , Keith Levin b , Qunying Huang a , Kenneth R. Mayer c (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-05636.txt.

Introduction

Jacob Kruse and co-authors extend the rich club idea to moving, place based networks in a 2025 paper. In many networks, the busiest nodes link strongly to each other. Measuring that club effect over time and space at once was missing. Their new metric weighs link strength and steadiness across time steps.

A unified frame lets the club variants be measured side by side. Two human movement case studies test the metric at different scales. The weighted time aware version found real club effects the plain version missed. Maps from the analysis show where and when the clubs hold.

Catalog record

CREATORJacob Kruse a , Song Gao a,* , Yuhan Ji a , Keith Levin b , Qunying Huang a , Kenneth R. Mayer c
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE67 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-05636

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Identifying rich clubs in spatiotemporal interaction networks Jacob Kruse a , Song Gao a,* , Yuhan Ji a , Keith Levin b , Qunying Huang a , Kenneth R. Mayer c Abstract Spatial networks are widely used in various fields to represent and analyze interactions or relationships between locations or spatially distributed entities. While existing studies have proposed methods for hub identification and community detection in spatial networks, relatively few have focused on quantifying the strength or density of connections shared within a community of hubs across space and time. Borrowing from network science, there is a relevant concept known as the ‘rich club’ phenomenon, which describes the tendency of ‘rich’ nodes to form densely interconnected sub-networks. Although there are established methods to quantify topological, weighted, and temporal rich clubs individually, there is limited research on measuring the rich club effect in spatially-weighted temporal networks, which could be particularly useful for studying dynamic spatial interaction networks. To address this gap, we introduce the spatially-weighted temporal rich club (WTRC), a metric that quantifies the strength and consistency of connections between rich nodes in a spatiotemporal network. Additionally, we present a unified rich club framework that distinguishes the WTRC effect from other rich club effects, providing a way to measure topological, weighted, and temporal rich club effects together. Through two case studies of human mobility networks at different spatial scales, we demonstrate how the WTRC is able to identify significant weighted temporal rich club effects, whereas the unweighted equivalent in the same network either fails to detect a rich club effect or inaccurately estimates its significance. In each case study, we explore the spatial layout and temporal variations revealed by the WTRC analysis, showcasing its particular value in studying spatiotemporal interaction networks. This research offers new insights into the study of spatiotemporal networks, with critical implications for applications such as transportation, redistricting, and epidemiology. keywords human mobility, spatial networks, spatiotemporal networks, spatial interactions, rich club † † affiliation: a Department of Geography, University of Wisconsin-Madison, USA b Department of Statistics, University of Wisconsin-Madison, USA c Department of Political Science, University of Wisconsin-Madison, USA † † thanks: *Corresponding author: [email protected] 1 Introduction Spatial networks are widely used across various fields, such as in geography, transportation, public health, environment and ecology, to represent and analyze connections and flows between locations or spatially distributed objects or entities ( Barthélemy, 2011 ; Gersmehl, 1970 ; Bian and Elwood, 2024 ) . In such networks, the edge weights encode the volume of flows or interactions between nodes (representing locations, spaces, and places), reflecting the effects of geographic distance, adjacency, and other processes on the strength or likelihood of connections between two locations or geographical entities ( Roy and Thill, 2004 ; Zhong et al., 2014 ; Andris, 2016 ; Clarke and Birkin, 2018 ; Xu et al., 2022 ) . One common use for spatial interaction networks is to model human mobility flows, and research across various disciplines has developed and utilized network methods to analyze human mobility flows in diverse ways. For instance, community detection methods have been applied to identify spatially connected geographic communities within these networks based on interaction strengths ( Gao et al., 2013 ; Chen et al., 2015 ; Liang et al., 2022 ; Wang et al., 2021 ) . Recent studies have extended these approaches to detect dynamic communities in spatiotemporal interaction networks ( Kang et al., 2022 ; Jia et al., 2022 ; Zhao et al., 2023 ) , revealing patterns in human mobility dynamics over time. In another line of research, patterns in human mobility flows and their changes in response to disruptions such as COVID-19 have been analyzed to understand spatiotemporal phenomena like disease transmission. For example, Chin et al., 2024 examined how urban mobility dynamics in one city evolved during the COVID-19 pandemic, while Kang et al., 2020 and Hou et al., 2021 analyzed how human mobility flows changed in response to COVID-19 infection at different geographic scales and over time. Since human mobility flow datasets are typically aggregated to geographic units; the nodes in such spatial interaction networks represent statistical areas rather than distinct entities directly corresponding to cities, towns, or other geographic features. The scale and specific boundary locations of these aggregated regions can significantly influence subsequent analyses—a challenge known as the modifiable areal unit problem (MAUP) ( Openshaw, 1984 ; Atkinson and Tate, 2000 ) . One common approach to addressing the MAUP problem is to perform analyses at multiple spatial scales, which helps to reveal how patterns change across different levels of aggregation and reflect the varying spatial processes at work ( Oshan et al., 2022 ; Quattrochi and Goodchild, 2023 ) . Given the additional complexity of analyzing spatial interaction patterns across scales, such studies often represent human mobility flows as undirected spatial networks. A relatively new application of human mobility network is in redistricting tasks. Redistricting is a type of regionalization task that involves dividing a larger geographic area into smaller discrete regions, often with specific goals such as equalizing population or ensuring fair representation. Recent studies have utilized human mobility networks to produce and evaluate district plans based on how well they align with underlying human mobility patterns across geographic communities. For example, Liang et al., 2022 ; Liang et al., 2025 and Wang et al., 2021 used human mobility networks to design health service areas that account for ease of access and existing mobility patterns. In the context of political redistricting, Kruse et al., 2024 evaluated district plans based on spatial interaction networks, with plans having relatively more intra-district flows considered superior to those having higher inter-district flows. One aspect of political redistricting research that remains under-explored is the use of dynamic human mobility flows to understand district cores. The preservation of district cores is a frequently cited but ambiguously defined criterion in redistricting research ( Eckman, 2021 ; Yablon, 2022 ) . A quantitative framework based on dynamic human movements could provide valuable insights into defining and understanding these core regions as spatial interaction communities, which can also contribute to the geographic process-oriented regionalization based on spatiotemporal data ( Zhang et al., 2024 ) . To analyze district cores using human mobility flows, it is necessary to employ a method capable of identifying core regions within the dynamic human mobility network (i.e., spatial connections or interaction strengths change over time). One promising concept from network science, known as the ‘rich club’ phenomenon ( Colizza et al., 2006 ) , could be particularly useful for identifying district cores in spatial interaction networks. Unlike community detection, which partition the entire (spatial) network to cohesive sub-networks ( Newman and Girvan, 2004 ) , the rich club phenomenon describes a network property that quantifies the tendency of ‘rich’ nodes—those with high connectivity or influence—to form densely interconnected sub-networks. In this context, rich club analysis can serve as a method in geographic networks to quantify and evaluate the strength of connections or flows among a subset of influential nodes within the networked geographic phenomena such as population migration ( Bian and Elwood, 2024 ; Koylu and Kasakoff, 2022 ) . Various definitions of the rich club exist ( Zhou and Mondragon, 2004 ;

Read the complete text