Mapping the Spatial Distribution of Poverty in Maldives

This report analyzes the potential benefits and challenges of machine learning and small area estimation (SAE) to monitor poverty and help tackle inequality in Maldives.

It explains how SAE can generate granular poverty estimates using household survey and census data, and how machine learning including convolutional neural networks were developed using a model trained on Indonesian data. The report highlights inconsistencies and challenges such as limited sample sizes and shows how developing localized models and synchronizing data collection could help Maldives improve its poverty mapping to drive equitable and targeted interventions.