Hafsah Fatihul Ilmy

Urbanization presents opportunities and challenges, with slum areas being a critical concern in many rapidly growing cities. Accurate identification and prediction of slum areas are imperative for effective urban planning and resource allocation. We propose using OpenStreetMap (OSM) Building data to identify slum areas in regions where slum datasets are lacking. The proposed methodology incorporates variable extraction using the "momepy" library to identify and characterize buildings effectively. In this study, we employ a two-stage process. Firstly, we utilize the momepy library to extract pertinent variables that encapsulate geometric and topological attributes of buildings from OSM data. These variables serve as inputs to our machine-learning model. Secondly, a predictive model is developed using machine learning algorithms for regions with known slum areas. This model is then employed to predict potential slum areas in regions with insufficient or lacking slum-related data. Preliminary experiments conducted on real-world datasets demonstrate the effectiveness of our approach. The results underscore the utility of OSM data in slum area detection, showcasing its capacity to bridge data gaps and facilitate proactive urban planning. Furthermore, the proposed approach offers adaptability to various urban contexts, making it a valuable tool for decision-makers and researchers in diverse regions. Additionally, the study contributes to ongoing discussions regarding the reliability of OSM data, especially in challenging urban environments like slums.