Spatial and Socioeconomic Drivers of Land Use Change in High-Density Residential Areas (R-3) in Quezon City: A Predictive Simulation Using the Cellular Automata-Markov Chain Model

Abstract. High-density residential (R-3) areas, characterized by flexible housing typologies and increasing population pressures, are particularly vulnerable to rapid urbanization, especially in highly urbanized cities like Quezon City. However, the complex interplay of spatial and socioeconomic factors driving R-3 zone changes is often overlooked in existing literature. To address this gap, this study employed binary logistic regression (BLR) to determine key drivers of R-3 zone transitions and the Cellular Automata (CA)–Markov model to predict future land use maps, utilizing historical data from 2000, 2009, and 2011. The BLR model, with acceptable discriminative ability (AUC = 0.78; pseudo-R² = 0.083), identified 24 statistically significant driving factors at the 95% confidence level. Results revealed that proximity to low-intensity industrial areas, central business districts, roads, informal settlements, vacant lots, population density, and specific soil types positively influenced R-3 transitions. Conversely, elevation, medium-intensity industrial areas, open spaces, cemeteries, reservoirs, waterways, and other soil types affect R-3 changes negatively, with distance from socialized housing and high traffic volume emerging as strong negative drivers. The suitability map, derived from the BLR model applied with a 5 x 5 neighborhood, and the transition areas matrix that accounts for the pixel-level land use shifts from the Markov model successfully projected a 2011 land use map with an acceptable model performance (Overall Kappa = 0.70). By identifying the statistically significant spatial and socioeconomic drivers influencing R-3 changes and predicting future developments, this study offers planners and local governments insights to address overcrowding, housing shortages, and informal settlement growth.