Spatial statistical urban-growth models of rapidly growing cities provide an effective means for identifying determinants of urban growth and generating projections of future growth. We set out to identify two specific types of growth – infill and expansion – and quantify their relationship with driving factors. For this purpose, land-cover changes were extracted from multi-temporal Landsat images for the period 1999 – 2017. Spatio-temporal urban growth processes and its types were identified, which allows the construction of spatial logistic regression (SLR) models and subsequently urban growth projections. Results indicate that population density, proportion of built-up land area in the surrounding area and land value are the strongest predictors of overall urban growth. For infill growth, the most significant predictors are distance to urban centre, proportion of built-up land area in the surrounding area, distance to major roads and population density, whereas land value and proportion of built-up land area in the surrounding area are the major predictors for expansion. Results show that the SLR models could not fully explain scattered, outward urban expansion. Modelling driving factors of urban growth types improves knowledge of urban growth patterns and processes; such knowledge is most relevant in support of urban planning and policy development.