Multi-Pattern Characteristics and Driving Mechanisms of Mixed Land Use: A Case Study of Changsha’s Built-Up Areas, China

Mixed land use (MLU) plays a pivotal role in promoting compact and sustainable urban development. This study examines Changsha’s built-up areas using the Place2vec model and Point of Interest (POI) data to identify MLU patterns. We quantitatively assess the degree of MLU through three dimensions: diversity, compatibility, and accessibility, and employ the Optimal Parameters-Based Geographical Detector (OPGD) model to uncover the driving factors influencing the MLU degree across different patterns. The results show that the Place2vec model identifies six mixed patterns in Changsha’s built-up areas, each encompassing diverse urban land types. The six mixed patterns exhibit significant differences in terms of diversity, compatibility, and accessibility, with the overall degree of MLU presenting a central–peripheral spatial structure. Although driving mechanisms vary across patterns, transport network connectivity and spatial utilization efficiency consistently exhibit dominant influences. These findings offer valuable insights for designing targeted urban planning strategies and optimizing land-use configurations to promote urban development.