Mapping fine-scale socioeconomic inequality using machine learning and remotely sensed data.

Area of Science:

• Socioeconomic research
• Geospatial analysis
• Data science

Background:

• Estimating socioeconomic inequality at fine scales is challenging due to limited data, particularly in lower- and middle-income countries.
• Existing methods struggle with spatiotemporal consistency and jurisdictional granularity.

Purpose of the Study:

• To develop and validate a novel data harmonization method for generating fine-scale socioeconomic inequality estimates.
• To improve the accuracy and granularity of inequality measurement in India.

Main Methods:

• Combined household survey data with freely available remote sensing data (including nighttime luminosity).
• Employed machine learning techniques and a novel data harmonization approach.
• Utilized a spatially cross-validated machine learning model with Demographic and Health Survey data.

Main Results:

• Achieved >84% prediction accuracy in estimating socioeconomic inequality measures.
• Identified key remote sensing datasets that enhance predictive power for inequality estimation.
• Demonstrated a reliable method for harmonizing asset and sociodemographic information.

Conclusions:

• The replicable approach effectively addresses data gaps in socioeconomic inequality at subnational levels.
• This method has the potential to enhance global inequality data for research and policy.
• Supports Sustainable Development Goals initiatives focused on reducing socioeconomic disparities.