Affected by the development strategy of Rural Space Reconstruction in China, the demand for rural mass housing has peaked in the North China Plain in the past 20 years. However, due to the inefficiency of conventional design methods, the rural houses built appear to have a noticeable trend of urbanization and homogeneity. To propose a more effective design approach to change the hitherto unsuccessful homogenized phenomenon of rural design, the study is based on investigating the composition, configuration and characteristics of the dwellings in some traditional villages of the eastern Shandong Province, and it compares and analyzes the differences between conventional methods and generation methods through three design tests: Test 1 is for the reappearance of a general mode of planning, Test 2 is based on the definition of shape-grammar-based rules and Test 3 is mainly used for the optimizing and programming of rules. Furthermore, based on the three prototypes of homestead combination, three-level rules are determined through the three tests mentioned above: Level-1 describes the housing prototype consisting of four homesteads, which generates a variety of spatial relations through the translation of homesteads. Level-2 describes a neighborhood prototype consisting of 16 homesteads, which generates various samples through splitting prototypes and expanding homesteads. Level-3 describes a block prototype consisting of 64 homesteads, which controls open space and identifies a given base during sample filling. Through the analysis of the tests results, the rationality and feasibility of the generative design approach are verified, proving that this approach effectively solves the design monotony problem that commonly exists in rural mass housing in the North China Plain.

Urban recreational spaces (URSs) are pivotal for enhancing resident well-being, making the accurate assessment of public perceptions crucial for quality optimization. Compared to traditional surveys, social media data provide a scalable means for multi-dimensional perception assessment. However, existing studies predominantly rely on single-modal data, which limits the comprehensive capturing of complex perceptions and lacks interpretability. To address these gaps, this study employs cutting-edge large vision–language models (LVLMs) and develops an interpretable model, Qwen2.5-VL-7B-SFT, through supervised fine-tuning on a manually annotated dataset. The model integrates visual-linguistic features to assess four perceptual dimensions of URSs: esthetics, attractiveness, cultural significance, and restorativeness. Crucially, we generate textual evidence for our judgments by identifying the key spatial elements and emotional characteristics associated with specific perceptions. By integrating multi-source built environment data with Optuna-optimized machine learning and SHAP analysis, we further decipher the nonlinear relationships between built environment variables and perceptual outcomes. The results are as follows: (1) Interpretable LVLMs are highly effective for urban spatial perception research. (2) URSs within Beijing’s Third Ring Road fall into four typologies, historical heritage, commercial entertainment, ecological-natural, and cultural spaces, with significant correlations observed between physical elements and emotional responses. (3) Historical heritage accessibility and POI density are identified as key predictors of public perception. Positive perception significantly improves when a block’s POI functional density exceeds 4000 units/km² or when its 500 m radius encompasses more than four historical heritage sites. Our methodology enables precise quantification of multidimensional URS perceptions, links built environment elements to perceptual mechanisms, and provides actionable insights for urban planning.