This study aims to identify an appropriate spatial analysis unit for the analysis and impact assessment of pedestrian traffic. It tries to address the limitations of previous studies that arbitrarily selected spatial units, by systematically deriving a spatial analysis unit optimized for pedestrian traffic analysis. Method: The study established 50×50 meters as the minimum spatial analysis unit and, based on this, constructed pedestrian traffic datasets for various spatial scales, including 100×100m, 200×200m, 400×400m, and 500×500m. Using these datasets, histogram analysis, spatial autocorrelation analysis (Moran’s I), correlation analysis, and simple regression analysis were conducted to determine the optimal spatial unit. In addition, commercial floor area data were similarly constructed and analyzed through correlation and regression analyses with the pedestrian traffic data. The data preprocessing stage also examined the influence of including or excluding zero values. Result: The analysis revealed that both pedestrian traffic and commercial floor area data exhibited clear density differentiation starting from the 100×100m unit, also spatial autocorrelation, correlation, and regression analyses showed the most stable results from this spatial unit. In particular, the 100×100m unit was confirmed to be the most suitable for pedestrian traffic analysis, although the 200×200m unit could also be used with caution due to potential information loss caused by smoothing effects. This study provides a practical guideline for setting spatial units in pedestrian traffic analysis and suggests directions for further research.