Mapping The Intellectual Landscape of Stunting Prediction Using Satellite Imagery and Machine Learning: A Bibliometric Analysis

Research on stunting risk prediction has become increasingly interdisciplinary, integrating demographic, environmental, and computational approaches. This study maps the global research landscape on stunting prediction using demographic data, satellite imagery, and machine learning, employing a hybrid methodology that combines a systematic scoping review and bibliometric analysis. Data from Scopus and Web of Science (248 publications, 2009–2025) were analyzed using Bibliometrix to examine publication trends, thematic structures, and collaboration patterns. Results show rapid growth after 2019, driven by the availability of demographic and remote-sensing data and advances in machine learning. Two dominant clusters were identified: a health–data cluster (machine learning, public health, malnutrition) and a socio-geospatial cluster (geography, remote sensing, poverty, environmental health). Thematic mapping indicates that medicine remains a basic theme, geography functions as a motor theme, and machine learning occupies a transitional position signifying ongoing methodological development. The scoping synthesis reveals a clear shift toward data-driven, spatially explicit research that integrates socioeconomic and environmental factors. However, gaps remain in model generalization, cross-regional comparison, and the inclusion of behavioral and climatic dimensions. This study provides the first comprehensive hybrid mapping of the field, illustrating its transition from health-based analyses to an interdisciplinary, data-science-driven paradigm. The findings offer a roadmap for researchers and policymakers to enhance collaboration, methodological rigor, and evidence-based actions aligned with Sustainable Development Goal 2 (Zero Hunger).