Correlation analysis between pathfinding factors and spatial attributes

Table 1: Wayfinding Performance at Key Nodes
Core content: Average errors and error rates at 8 key nodes of Xizhimen Station. Variables: Node ID (unique identifier), Average Number of Errors (mean errors per node across 30 participants), Wayfinding Error Rate (calculated as [Average Errors/(1+Average Errors)]×100). Key feature: Errors concentrated at Node 3 (25.00%), 4 (37.50%), 6 (40.00%); others show 0% error.
Table 2: Spatial Attributes and Wayfinding Performance
Core content: Integrates independent variables (Spatial Connectivity/Visual Area) and dependent variables (Avg. Decision Time/Error Rate) for 8 nodes. Variables: Spatial Connectivity (number of directly connected units), Visual Area (m², isovist polygon area), Avg. Decision Time (s), Error Rate (%). Key feature: High connectivity/visual area nodes (4,6) correlate with higher error rates and longer decision times.
Table 3: Correlation (Error Rate vs. Spatial Attributes)
Core content: Pearson correlation results (n=8, two-tailed α=0.05). Results: SC-ER (r=0.411, p=0.311), VA-ER (r=0.414, p=0.308) — weak positive correlations, not statistically significant due to small node sample.
Table 4: Correlation (Decision Time vs. Spatial Attributes)
Core content: Pearson correlation results (n=8, two-tailed α=0.05). Results: SC-DT (r=0.513, p=0.194), VA-DT (r=0.510, p=0.197) — moderate positive correlations, near significance.
Table 5: Grouped Analysis (Spatial Connectivity)
Core content: Detailed SC, DT, ER for each node. Role: Input data for SC-DT/SC-ER cubic regression equations.
Table 6: Predicted vs. Measured ER (Spatial Connectivity)
Core content: Compares measured ER, regression-predicted ER, and residuals. Key validation: High accuracy (e.g., Node 6: predicted 40.12%, measured 40.00%), confirming model reliability.
Table 7: Grouped Analysis (Visual Area)
Core content: Detailed VA, DT, ER for each node. Role: Input data for VA-DT/VA-ER cubic regression equations.
Table 8: Predicted vs. Measured ER (Visual Area)
Core content: Compares measured ER, regression-predicted ER, and residuals. Key validation: Low prediction deviation, supporting nonlinear relationship validity.
Table 9: Visual Area Safety Threshold
Core content: Classifies nodes as "Safe" (ER≤20%)/ "Risk" (ER>20%) based on 538–955 m² threshold. Key criterion: Nodes within threshold are mostly safe (except Node 3), verifying threshold effectiveness.