A framework to determine micro-level population figures using spatially disaggregated population estimates

TL;DR

This paper presents a framework utilizing high-resolution spatial data, GIS tools, and machine learning to generate micro-level population estimates, improving urban planning and resource allocation especially in data-scarce regions.

Contribution

It introduces a novel framework combining spatial data and machine learning to estimate local population figures, addressing data gaps in developing countries.

Findings

Accurately estimated population at micro levels in Lagos Island, Nigeria.

Demonstrated framework's potential for urban service capacity planning.

Validated the approach as a benchmark for local population estimation.

Abstract

About half of the world population already live in urban areas. It is projected that by 2050, approximately 70% of the world population will live in cities. In addition to this, most developing countries do not have reliable population census figures, and periodic population censuses are extremely resource expensive. In Africa's most populous country, Nigeria, for instance, the last decennial census was conducted in 2006. The relevance of near-accurate population figures at the local levels cannot be overemphasized for a broad range of applications by government agencies and non-governmental organizations, including the planning and delivery of services, estimating populations at risk of hazards or infectious diseases, and disaster relief operations. Using GRID3 (Geo-Referenced Infrastructure and Demographic Data for Development) high-resolution spatially disaggregated population data estimates, this study proposed a framework for aggregating population figures at micro levels within a larger geographic jurisdiction. Python, QGIS, and machine learning techniques were used for data visualization, spatial analysis, and zonal statistics. Lagos Island, Nigeria was used as a case study to demonstrate how to obtain a more precise population estimate at the lowest administrative jurisdiction and eliminate ambiguity caused by antithetical parameters in the calculations. We also demonstrated how the framework can be used as a benchmark for estimating the carrying capacities of urban basic services like healthcare, housing, sanitary facilities, education, water etc. The proposed framework would help urban planners and government agencies to plan and manage cities better using more accurate data.