A quantitative and direct evaluation method of walking environment using location-based big data

Traditional walkability metrics, such as the Highway Capacity Manual (HCM) or static GIS-based scores, typically evaluate physical capacity based on absolute density or static environmental factors. However, these methods fail to capture the dynamic temporal variations and the specific “conflict” (intersection) of pedestrian flows that affect the walking experience. This study proposes a novel, dynamic walkability evaluation method using location-based big data. We introduce a “Walkability Index” calculated from the product of normalized Z -scores for inflow and outflow vectors within a mesh. This mathematical design is intentionally structured to detect flow intersection—where multidirectional movements occur simultaneously—rather than simple volume. We applied this method to two distinct urban areas in Saitama, Japan: Omiya Station (a major transport hub) and Kawagoe City (a tourist site). The statistical analysis revealed that in the Omiya area, the third quartile (Q3) of Z -scores remained below 0.1 across all time periods. This indicates that “low walkability” is not a uniform condition but is highly localized in specific statistical outliers (hotspots) representing the top 25% of meshes. In contrast, Kawagoe exhibited linear distribution patterns along tourist routes. While full ground-truth validation remains a future task, this method offers a scalable, quantitative tool for city planners to pinpoint specific dynamic bottlenecks, thereby complementing traditional static evaluations for targeted interventions.