Discovering the Spread Patterns of SARS-CoV-2 in Metropolitan Areas

Herrera Marín, Mauricio René; Vergara Perucich, Francisco; Aguirre-Núñez, Carlos; Godoy-Faúndez, Alex

Publication:
Discovering the Spread Patterns of SARS-CoV-2 in Metropolitan Areas

dc.contributor.author	Herrera Marín, Mauricio René
dc.contributor.author	Vergara Perucich, Francisco
dc.contributor.author	Aguirre-Núñez, Carlos
dc.contributor.author	Godoy-Faúndez, Alex
dc.date.accessioned	2024-04-16T20:27:02Z
dc.date.available	2024-04-16T20:27:02Z
dc.date.issued	2023
dc.description.abstract	The spread of COVID-19 has been extensively studied, but the intricate dynamics of its transmission in interdependent and segregated urban areas, constrained by mobility restrictions, have not been completely understood yet. The pandemic's dynamic-adaptive nature implies that virus spread is influenced by diverse factors operating disparately in urban areas with distinct roles. This study investigates the dynamic spread patterns of COVID-19 in the Santiago Metropolitan Area (SMA), Chile, leveraging explanatory variables related to urban mobility, socio-spatial characteristics, segregation, and sanitary measures. Using publicly available mobility data, we used two indices—the Internal Mobility Index (capturing individual trips within a city’s commune), and the External Mobility Index (indicating trips crossing commune borders). These indices were derived from geolocation data recorded by the cellular telephone antenna network of the Telefónica company by tracking successive antenna transitions during trips. The analysis encompasses a three-stage pandemic pattern, corresponding to periods before, during, and after an initial lockdown in the pandemic's first year. Elastic-Net-Penalty regression models, skillful in both feature selection and managing highly correlated predictors while maintaining the interpretability of the models, are used. These models employ a combination of L1 (ridge) and L2 (lasso) regularized log-likelihood optimization. The ridge penalty functions by contracting the coefficients of correlated predictors, pulling them closer to each other. In contrast, the lasso method tends to choose one predictor and exclude the others. The analysis with these models unveils influences of various explanatory variable subsets throughout the pandemic. Importantly, the study provides evidence justifying the suboptimal outcomes of the dynamic quarantine imposed by authorities. Mobility restrictions were implemented without considering the intricate contextual factors, thus impacting vulnerable areas of the city adversely.
dc.description.version	Versión publicada	es
dc.format.extent	21 p.	es
dc.identifier.citation	M.-R. Herrera-Marín, F. Vergara-Perucich, C. Aguirre-Núñez, A. Godoy-Faúndez, “Discovering the Spread Patterns of SARS-CoV-2 in Metropolitan Areas,” Revista Facultad de Ingeniería, vol. 32, no. 66, e16457, 2023.
dc.identifier.doi	10.19053/01211129.v32.n66.2023.16457	es
dc.identifier.issn	0121-1129
dc.identifier.uri	https://hdl.handle.net/11447/8631
dc.language.iso	en
dc.rights	Atribución-NoComercial-CompartirIgual 3.0 Chile (CC BY-NC-SA 3.0 CL)
dc.rights.uri	http://creativecommons.org/licenses/by-nc-sa/3.0/cl/
dc.subject	city dynamics	en
dc.subject	COVID-19 spread dynamics	en
dc.subject	elastic-net regularization	es
dc.subject	socio-spatial mobility indicators	en
dc.subject	time-varying regression	en
dc.title	Discovering the Spread Patterns of SARS-CoV-2 in Metropolitan Areas
dc.type	Article
dcterms.accessRights	Acceso abierto	es
dcterms.source	Revista Facultad de Ingeniería	es
dspace.entity.type	Publication