Identifying deprived "slum" neighbourhoods in the Greater Accra Metropolitan Area of Ghana using census and remote sensing data

Robert MacTavish; Honor Bixby; Alicia Cavanaugh; Samuel Agyei-Mensah; Ayaga Bawah; George Owusu; Majid Ezzati; Raphael Arku; Brian Robinson; Alexandra M Schmidt; Jill Baumgartner

doi:10.1016/j.worlddev.2023.106253

Identifying deprived "slum" neighbourhoods in the Greater Accra Metropolitan Area of Ghana using census and remote sensing data

World Dev. 2023 Jul:167:106253. doi: 10.1016/j.worlddev.2023.106253.

Authors

Robert MacTavish^{1

2}, Honor Bixby^{1

2

3}, Alicia Cavanaugh⁴, Samuel Agyei-Mensah⁵, Ayaga Bawah⁵, George Owusu^{5

6}, Majid Ezzati⁷, Raphael Arku^{8

9}, Brian Robinson⁴, Alexandra M Schmidt¹, Jill Baumgartner^{1

2}

Affiliations

¹ Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Canada.
² Institute for Health and Social Policy, McGill University, Montreal, Canada.
³ Institute of Public Health and Wellbeing, University of Essex, Colchester, England.
⁴ Department of Geography, McGill University, Montreal, Canada.
⁵ Department of Geography and Resource Development, University of Ghana, Accra, Ghana.
⁶ Institute of Statistical, Social and Economic Research, University of Ghana, Accra, Ghana.
⁷ Faculty of Medicine, School of Public Health, Imperial College, London, England.
⁸ Institute for Global Health, University of Massachusetts Amherst, Amherst, United States.
⁹ Department of Environmental Health Sciences, University of Massachusetts Amherst, Amherst, United States.

Abstract

Background: Identifying urban deprived areas, including slums, can facilitate more targeted planning and development policies in cities to reduce socio-economic and health inequities, but methods to identify them are often ad-hoc, resource intensive, and cannot keep pace with rapidly urbanizing communities.

Objectives: We apply a spatial modelling approach to identify census enumeration areas (EAs) in the Greater Accra Metropolitan Area (GAMA) of Ghana with a high probability of being a deprived area using publicly available census and remote sensing data.

Methods: We obtained United Nations (UN) supported field mapping data that identified deprived "slum" areas in Accra's urban core, data on housing and population conditions from the most recent census, and remotely sensed data on environmental conditions in the GAMA. We first fitted a Bayesian logistic regression model on the data in Accra's urban core (n=2,414 EAs) that estimated the relationship between housing, population, and environmental predictors and being a deprived area according to the UN's deprived area assessment. Using these relationships, we predicted the probability of being a deprived area for each of the 4,615 urban EAs in GAMA.

Results: 899 (19%) of the 4,615 urban EAs in GAMA, with an estimated 745,714 residents (22% of its urban population), had a high predicted probability (≥80%) of being a deprived area. These deprived EAs were dispersed across GAMA and relatively heterogeneous in their housing and environmental conditions, but shared some common features including a higher population density, lower elevation and vegetation abundance, and less access to indoor piped water and sanitation.

Conclusion: Our approach using ubiquitously available administrative and satellite data can be used to identify deprived neighbourhoods where interventions are warranted to improve living conditions, and track progress in achieving the Sustainable Development Goals aiming to reduce the population living in unsafe or vulnerable human settlements.

Keywords: Informal settlements; Satellite imagery; Urban poverty.

Abstract

Grants and funding