Abstract: The phenomenon of shared living, characterized by the high integration of spatialstructures and social relations, represents an efficient model of neighborhood residence and is oneof the challenges in the renewal of residential historic areas. From both spatial and socialdimensions, with the smallest residential unit as a node, this study identifies multiple connectiontypes at the urban, shared, and private levels based on depth and path methods, and constructs aspatial network using spatial gravity. Simultaneously, it constructs an adjacency matrix reflectingrelationships among acquaintances, semi-acquaintances, and strangers to form a weightednetwork with social characteristics. By combining the two through matrix operations, a "relational-spatial" network is developed to reveal the comprehensive structural characteristics and unitcluster features of residential historic areas, deriving a unit-based identification method for sharedliving. The validity of this identification method is tested with samples under different ownershipstatuses, and the scientific and practical value of the "relational-spatial" network and unit-basedidentification method are discussed in terms of spatial cognition, renewal design, and institutional
practice in historic areas.