We've Moved!
Visit SDSU’s new digital collections website at https://digitalcollections.sdsu.edu
Description
This study explored the connection between remotely sensed imagery and ground-based housing and welfare survey data in slum neighborhoods in Accra, Ghana. Specific household-level variables reflective of housing quality and demographics from the 2009-2010 Housing and Welfare Study (HAWS) of Accra and the 2003 UN-Habitat Accra Slum Survey (AccraSS) were regressed against measures extracted from high spatial resolution Quickbird satellite imagery captured in 2002 and again in 2010. Samples from the two surveys for 37 census enumeration areas (EAs) within the Accra Metropolitan Area (AMA) were analyzed. An exhaustive regression analysis was run to measure the covariation between individual survey data variables and metrics derived from the imagery. A spatial regimes approach explored spatially autocorrelated data in the discontinuous data set, and these results were compared with a geographically weighted regression approach. The goal was to establish "proxy" variables from satellite remote sensing data that are indicative of household health and welfare characteristics over time by combining spatially homogenous predictors in multivariate regression models. By generating proxies of the built environment, we may be able to infer or extrapolate socioeconomic and health statuses for each respective EA and the surrounding neighborhoods at other dates (e.g., between surveys and censuses). Specifically, I test the hypotheses that (1) socioeconomic and demographic characteristics of slum areas can be inferred from spatial variations in vegetation and texture as derived from satellite imagery; and (2) dynamics of socioeconomic and demographic characteristics can be quantified from changes in the image metrics. Since one in six residents of the world is estimated by the UN to be living in a slum, it is important to understand how these neighborhoods might transform over time as the population of a major city in a developing nation increases at a high rate.
