Using comparative approaches to model deprivation in Antananarivo, Madagascar: A multidimensional analysis using principal components analysis and weighting system across meso and macro scales

Abstract

Rapid population growth and global urbanization pose socio-economic challenges, causing inequalities in Global South (GS) cities, such as the proliferation of deprived areas. This study aims to develop methods for mapping and characterizing urban deprivation in GS cities using the IDEAMAPS framework, focusing on household, area, and area-connect levels. The methodology employs 23 indicators across five domains, with a weighting system applied at macro (agglomeration of Antananarivo) and meso (Urban Commune of Antananarivo - CUA) scales, alongside Principal Component Analysis (PCA) and population-weighted analysis. A variable reduction process assessed the impact of simplifying indicators while retaining explanatory power. Results demonstrated significant spatial contrasts in deprivation between central and peripheral areas and eastern and western neighborhoods. The equal weighting system provided an intuitive overview, showing that 53% of neighborhoods were privileged at the macro scale, while 15% were highly deprived. At the meso scale, 27% of neighborhoods were highly deprived, emphasizing the importance of finer spatial scales to uncover localized disparities. PCA reduced data complexity and identified key deprivation dimensions but remains sensitive to outliers. Population-weighted analysis revealed the misalignment between deprivation level and population density, highlighting the need for targeted interventions in densely populated neighborhoods.

Variable reduction confirmed model robustness but underscored the importance of retaining critical variables. This study highlights the need for accurate, multi-scale assessments to inform policies addressing urban inequalities. Future research should integrate advanced spatial techniques, temporal dynamics, and additional indicators, such as governance and environmental hazards, to refine deprivation analyses and guide inclusive urban policies.