Mapping seasonal human mobility across Africa using mobile phone location history and geospatial data

TL;DR

This study uses mobile phone data and advanced modeling to map seasonal human movement across Africa, improving understanding of mobility patterns in data-scarce regions.

Contribution

The study introduces a Bayesian spatiotemporal framework to estimate mobility in data-limited African countries using novel data sources.

Findings

Monthly mobility flows were estimated for 25 African countries using the Google Aggregated Mobility Research Dataset.

Regional models filled data gaps for 28 additional countries with sparse or missing records.

Key drivers of mobility included economic, demographic, and environmental factors like GDP and evapotranspiration.

Abstract

Seasonal human mobility data are essential for understanding socioeconomic and environmental dynamics, yet much of Africa lacks comprehensive mobility datasets. Human movement, shaped by economic needs, family responsibilities, seasonal climatic variations, and displacements, is poorly documented in many regions due to limitations of traditional methods like censuses and surveys. This study addresses these gaps by leveraging the Google Aggregated Mobility Research Dataset (GAMRD) and a Bayesian spatiotemporal framework to estimate pre-pandemic monthly mobility flows at both national and regional scales across Africa for 2018–2019. We analysed 25 countries with complete GAMRD data and developed regional models to estimate mobility in 28 additional countries with sparse or missing records, filling critical data gaps. Key predictors, including GDP per capita, underweight children, infant mortality, environmental variables like stream runoff and evapotranspiration, and covariate interactions, revealed the complexity of mobility drivers. This approach provides robust estimates of seasonal mobility changes in data-limited areas, and offers a foundational understanding of African mobility dynamics, which highlights the value of innovative modelling and novel sources to bridge data gaps for supporting regional planning and policy-making.

Figures