A global 1 km gridded dataset of household consumption expenditure in purchasing power parity 2013-2024.

Household and non-profit institutions serving households (NPISH) final consumption expenditure is widely used to characterize welfare, demand, and living standards, but official statistics are typically available at national and regional scales and therefore cannot support fine-grained analyses. Here we present a global 1 km gridded dataset of household and NPISH final consumption expenditure in purchasing power parity (PPP) for 2013-2024. The dataset integrates World Bank economy-level totals, expressed in constant 2021 international dollars, with annual nighttime lights, population, built-up surface, roads, and consumer-oriented points-of-interests (POIs) in a stacked Super Learner framework. Harmonized annual covariates are used to generate grid-level predictions of per-capita consumption. The dataset is a model-based spatial disaggregation constrained by economy totals. We release two products: an observed-only variant based exclusively on directly reported World Bank totals, and a regionally completed variant that fills missing economy totals from World Bank regional aggregates. Each product includes annual cell-total rasters, per-capita rasters, QA rasters, and an economy-year metadata table. Independent validation against official subnational consumption statistics in the United States, Canada, and the United Kingdom, together with a provincial retail sales proxy validation for China, shows that the dataset captures spatial heterogeneity while retaining economy-level consistency. The dataset is intended to support analyses of welfare geography, market access, and consumption-related human activity, including applications relevant to the Sustainable Development Goals (SDGs), such as poverty, equity and inclusive development.