Mashhoodi, Bardia ORCID: https://orcid.org/0000-0002-7037-3932, Dane, Gamze, Hu, Yang, Zhong, Jian ORCID: https://orcid.org/0000-0003-1026-8695, Che, Yue, Ge, Ying-En and Sharifi, Ayyoob (2026) Education and environmental inequality: multi-hazard, non-linear, and spatially varying. Sustainable Cities and Society (SCS), 151:107956. ISSN 2210-6707 (Print), 2210-6715 (Online) (doi:10.1016/j.scs.2026.107956)

Preview

PDF (Open Access Article) 54873 MASHHOODI_Education_And_Environmental_Inequality_(OA)_2026.pdf - Published Version Available under License Creative Commons Attribution. Download (8MB) | Preview

Abstract

Existing analyses on educational attainment and environmental inequalities are often single-hazard, linear and spatially stationary. To bridge these gaps, we study associations between the share of residents in neighbourhoods with low educational attainment and Noise pollution, PM₂.₅, NO₂, O₃, Physiological Equivalent Temperature (PET) simulated for a tropical day, and green surfaces in 1,467 urbanised neighbourhoods of the Netherlands, controlling for socioeconomic factors. For the country-scale associations, an aspatial XGBoost model (Global-XGBoost) produces robust goodness-of-fit but high overfitting, indicating location-specific variations in the associations. Accordingly, a complementary geographically weighted XGBoost (GW-XGBoost) model is developed on local associations. The results show that the importance of environmental variables is higher (22% compared to 11%) in the GW-XGBoost models than in Global-XGBoost, and two clusters of neighbourhoods with low and high environmental sensitivities, mainly concentrated inside or outside of highly urbanised areas, could be distinguished. The predictions are both non-linear and spatially varying. At the country scale, neighbourhoods’ lower educational attainment is associated with higher PM₂.₅, NO₂, and PET, lower Noise exposure, lower Non-tree and Tree coverages, and the association with O3 is U-shaped. The associations differ for NO₂ (declining), as well as O3 and Non-tree vegetated surfaces (increasing) at the local scale.

Item Type:	Article
Uncontrolled Keywords:	environmental justice, GeoAI, air quality, heatwave, geographically weighted machine learning
Subjects:	H Social Sciences > H Social Sciences (General) L Education > L Education (General) Q Science > Q Science (General) Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Faculty / School / Research Centre / Research Group:	Faculty of Engineering & Science > School of Computing & Mathematical Sciences (CMS) Faculty of Engineering & Science
Last Modified:	02 Nov 2026 16:40
URI:	https://gala.gre.ac.uk/id/eprint/54873

Actions (login required)

View Item

Downloads