Fig 1.
Research framework.
Fig 2.
The study area and the spatial distribution of research units.
The basemap, including roads and building outlines, is based on data from OpenStreetMap contributors (Open Database License, ODbL). Building outlines were further supplemented and corrected by the authors through field surveys. Colored research units, city group, water and green land were independently delineated and visualized by the authors in GIS according to the research objectives.
Fig 3.
Delineation process of research units based on streets and adjacent storefronts.
Fig 4.
Dianping rating dimensions and corresponding sub-indicators.
Table 1.
Independent variables and formulas.
Fig 5.
Network communities: (a) Urban morphology; (b) Urban functionality. Node size denotes betweenness centrality and edge thickness shows association strength. Key variables were selected based on centrality thresholds.
Fig 6.
Workflow for street view image collection and analysis.
Street view images are collected at sampled points, semantically segmented using PSPNet, and quantitatively analyzed for key visual environmental indicators.
Fig 7.
Spatiotemporal distribution of social vitality within the street markets of Shangxia Sha urban village.
Maps illustrate the spatial distribution of social vitality within the street markets during seven time periods: (a) morning; (b) noon peak; (c) afternoon; (d) evening peak; (e) night; (f) early morning; (g) full-day average. The basemap, including roads and building outlines, is based on data from OpenStreetMap contributors (Open Database License, ODbL). Building outlines were further supplemented and corrected by the authors through field surveys. Colored research units were created by the authors in GIS software.
Fig 8.
Spatial distribution of commercial vitality and built environment factors within the street markets of Shangxia sha urban village.
The basemap, including roads and building outlines, is based on data from OpenStreetMap contributors (Open Database License, ODbL). Building outlines were further supplemented and corrected by the authors through field surveys. Colored research units were created by the authors in GIS software.
Table 2.
Results of global spatial autocorrelation analysis.
Table 3.
Results of variable collinearity diagnostics.
Table 4.
Comparison of OLS, GWR, and GTWR fitting results.
Table 5.
Descriptive statistics of average GTWR coefficients estimation across time periods.
Fig 9.
Spatial distribution of GTWR coefficients for urban morphology indicators.
Each subfigure displays the GTWR coefficient values for key urban morphology indicators (building density, street length, street width, integration and choice) at six time periods: (a) morning, (b) noon peak, (c) afternoon, (d) evening peak, (e) night, and (f) early morning, revealing their spatial heterogeneity and temporal dynamics within the street markets. The basemap, including roads and building outlines, is based on data from OpenStreetMap contributors (Open Database License, ODbL). Building outlines were further supplemented and corrected by the authors through field surveys. Colored research units were created by the authors in GIS software.
Fig 10.
Spatial distribution of GTWR coefficients for urban functionality indicators.
Each subfigure displays the GTWR coefficient values for key urban functionality indicators (POI density, mixing degree, life service facility density, subway convenience, bus convenience, and development intensity) at six time periods: (a) morning, (b) noon peak, (c) afternoon, (d) evening peak, (e) night, and (f) early morning, revealing their spatial heterogeneity and temporal dynamics within the street markets. The basemap, including roads and building outlines, is based on data from OpenStreetMap contributors (Open Database License, ODbL). Building outlines were further supplemented and corrected by the authors through field surveys. Colored research units were created by the authors in GIS software.
Fig 11.
Spatial distribution of GTWR coefficients for human perception indicators.
Each subfigure displays the GTWR coefficient values for key human perception indicators (green vision rate, enclosure degree, walkability, and facility visibility) at six time periods: (a) morning, (b) noon peak, (c) afternoon, (d) evening peak, (e) night, and (f) early morning, revealing their spatial heterogeneity and temporal dynamics within the street markets. The basemap, including roads and building outlines, is based on data from OpenStreetMap contributors (Open Database License, ODbL). Building outlines were further supplemented and corrected by the authors through field surveys. Colored research units were created by the authors in GIS software.
Table 6.
Comparison of OLS, GWR, and MGWR fitting results.
Table 7.
Statistical analysis of regression results in the MGWR model.
Fig 12.
Spatial patterns of coefficients in the MGWR model.
Each panel shows the spatial patterns of MGWR coefficients for the following variables: intercept, street length, street width, integration, POI density, life service facility density, development density, enclosure degree, walkability, and green vision rate. The basemap, including roads and building outlines, is based on data from OpenStreetMap contributors (Open Database License, ODbL). Building outlines were further supplemented and corrected by the authors through field surveys. Colored research units were created by the authors in GIS software.