Urban Morphologic Structures Retrieved by Satellite Imagery Correlate With Socioeconomic Household Data—Insights From the City of Kigali, Rwanda

Abstract

A substantial body of research exists on the use of remote sensing in urban contexts. However, only a limited number of studies have contributed to our understanding of the socioeconomic conditions of different urban areas. This research aims to demonstrate the potential of very high-resolution images and geospatial data by examining the interrelations between socioeconomic data retrieved from household surveys in the city of Kigali and spatial data on urban morphology retrieved by satellite imagery. As the surveys yielded large amounts of data of varying levels of measurement (categorical and numeric), we present different methods of statistical correlation, data mining, and machine learning to highlight socioeconomic patterns within the spatial data. The results demonstrate a significant correlation between the share of different building types, building density, average building heights, and distances to public infrastructure with a range of surveyed data, including building properties, household members, financial resources, and overall lifestyle habits. This highlights the potential of remote sensing and geospatial data to provide valuable insights into the socioeconomic conditions of urban areas. It also underscores the importance of using advanced statistical methods, data mining, and machine learning to enhance our understanding of urban morphology and its socioeconomic implications. However, it is important to acknowledge the limitations of such approaches, including the lack of information on ownership, potential for false inference and the direction of causation, which require further investigation.

Keywords

• Calibration
• data mining
• demography
• scientific visualization
• social factors
• urban areas