This study explores the spatial directionality of urban mobility, a critical yet often overlooked aspect, using vector computation to introduce a two-dimensional anisotropy-centripetality metric. Utilizing travel data from 90 million mobile users across 60 Chinese cities, the research quantifies mobility patterns, distinguishing between strong monocentric, weak monocentric, and polycentric patterns. Key findings include: 1. **Monocentric Cities**: Residents face increasing commuting distances as cities expand, with mobility anisotropy intensifying in the outskirts. 2. **Polycentric Cities**: Commuting patterns remain consistent, with lower anisotropy and centripetality, indicating more isotropic and centrifugal behaviors. 3. **Anisotropy and Centripetality**: These metrics effectively capture the spatial directionality of mobility, with higher anisotropy and centripetality in monocentric cities and lower values in polycentric cities. 4. **Correlations**: The average commuting distance is positively correlated with city size in monocentric cities but not in polycentric cities, suggesting greater mobility efficiency in the latter. 5. **Spatial and Temporal Dynamics**: Anisotropy increases with spatial level in all city types, while centripetality decreases, reflecting a core-periphery structure. Temporal dynamics show significant fluctuations during morning peak hours, with strong monocentric cities exhibiting the most pronounced changes. 6. **Microscopic Model**: A random workplace and residence choice (RWRC) model replicates observed commuting characteristics, highlighting the interplay between employment attraction strength and commuting distance scale. These insights are crucial for urban planning and policy-making, providing a framework to manage congestion and guide suburban development. The study also challenges traditional classifications of urban forms and emphasizes the dynamic nature of urban mobility patterns.This study explores the spatial directionality of urban mobility, a critical yet often overlooked aspect, using vector computation to introduce a two-dimensional anisotropy-centripetality metric. Utilizing travel data from 90 million mobile users across 60 Chinese cities, the research quantifies mobility patterns, distinguishing between strong monocentric, weak monocentric, and polycentric patterns. Key findings include: 1. **Monocentric Cities**: Residents face increasing commuting distances as cities expand, with mobility anisotropy intensifying in the outskirts. 2. **Polycentric Cities**: Commuting patterns remain consistent, with lower anisotropy and centripetality, indicating more isotropic and centrifugal behaviors. 3. **Anisotropy and Centripetality**: These metrics effectively capture the spatial directionality of mobility, with higher anisotropy and centripetality in monocentric cities and lower values in polycentric cities. 4. **Correlations**: The average commuting distance is positively correlated with city size in monocentric cities but not in polycentric cities, suggesting greater mobility efficiency in the latter. 5. **Spatial and Temporal Dynamics**: Anisotropy increases with spatial level in all city types, while centripetality decreases, reflecting a core-periphery structure. Temporal dynamics show significant fluctuations during morning peak hours, with strong monocentric cities exhibiting the most pronounced changes. 6. **Microscopic Model**: A random workplace and residence choice (RWRC) model replicates observed commuting characteristics, highlighting the interplay between employment attraction strength and commuting distance scale. These insights are crucial for urban planning and policy-making, providing a framework to manage congestion and guide suburban development. The study also challenges traditional classifications of urban forms and emphasizes the dynamic nature of urban mobility patterns.