Bayesian Factor Analysis of a Unidimensional Urban Sprawl Index in Ibadan, Nigeria

Abstract

Urban sprawl has been a phenomenon on increase over the past years, people migrate into the core part of a major city, due to different factors like available housing, land availability, infrastructural development or population growth among others, making it a major cause of concern. Therefore, this study aimed at investigating and understanding the factors influencing the rate of sprawl in Ibadan, Oyo state, Nigeria. The research considered multivariate and Bayesian techniques to characterize and understand the sprawl for a given period. The multivariate techniques were Independent Component and the Principal Component Analysis (ICA & PCA) while, the Bayesian technique was Bayesian Factor Analysis (BFA). These three techniques deal with extraction of underlying factors affecting sprawl indices. The result showed that the BFA identified the built-up density, openness, frag-saturation, and proximity as the major factors with the loadings (posterior values): 1.000, 0.412, 0.423, & 0.657 respectively. The ICA reduced to three underlying factors that could have effect on four variables (fragmentation, built-up density, urban density, cohesion). The PCA shared similar result of identified variables’ reduction (built-up dinsity, urban density, frag-saturation, proximity, and cohesion) to three factors with the ICA. Thus, the BFA extracted a single (unidimentional) index by reduction of the factors gotten, thereby making it easier to compute. Therefore, the major elements of sprawl are built-up density, fragmentation, openness, and proximity which could all be related to the Intensity of use of the land area. Thus, Government should put in place these four factors for urban planning or policy making especially the land areas to be covered by buildings and roads.