为有效评价以公共交通为导向的开发模式（Transit-Oriented Development, TOD）的实施效果，以轨道交通与城市的协调程度为切入点，基于经典的节点-场所（Node-Place, NP）模型，引入客流作为第三维度，提出节点-场所-客流（Node-Place-Ridership, NPR）模型，从三维度定量评估轨道交通站点的发展水平和维度间协调程度。以厦门轨道交通站点为例，基于多源时空大数据，分别从三维度计算共33项基础指标值，并以各指标的全局莫兰指数为权重计算各维度指标的加权和，得到各维度归一化的价值，最终利用K-means方法对各站点进行聚类分析。结果显示，NP模型将车站分为6类，NPR模型则进一步结合客流指标重新调整车站类别，修正节点或场所价值单一指标在结果分类中的过大权重，15座车站的分类在工作日NPR模型中有所变化。此外，工作日厦门地铁中高客流站点集中于岛内，其中1号线站点场所价值偏高，2、3号线站点则多属于平衡类;岛外1号线站点多属于依赖类，各项指标值均偏低，2、3号线站点节点价值偏高。研究表明站点分类与城市建设进度紧密相关，站域范围内的交通、土地利用与客流之间存在普遍联系，在未来开发中应进一步有针对性地指导规划建设，促进轨道交通与城市融合发展。

In order to effectively evaluate the implementation effect of TOD (Transit-Oriented Development), taking the coordination degree between rail transit and cities as the entry point, the paper proposed the NPR (Node-Place-Ridership) model based on the classical NP (Node-Place) model and introduced ridership as the third dimension to quantitatively assess the development level of rail transit stations from three dimensions and the coordination among dimensions. Taking Xiamen Metro stations as an example, based on the spatio-temporal data with multi-source, the values of 33 basic indicators were calculated from the three dimensions, and the global Moran index of each indicator was used as the weight to calculate the weighted sum of each dimensional indicator to get the normalized value of each dimension. Finally, the K-means method was used to cluster the stations. The results showed that the NP model divided the stations into 6 categories, while the NPR model further readjusted the station categories in combination with the ridership indicator, and corrected the excessive weight of a single indicator of node or place value in the result classification, and the classification of 15 stations was changed in the weekday NPR model. In addition, on weekdays, the stations of Xiamen metro with medium-high ridership were concentrated in the island, of which the place value of Line 1 was relatively high, while the stations of Line 2 and Line 3 mostly belonged to the balance category. Meanwhile, off-island stations of Line 1 mostly belonged to the dependency category, with low indicator values, and stations of Line 2 and Line 3 had higher node value. The research demonstrates that the classification of stations is closely related to the progress of urban construction, and there is a general connection between the transportation, land use and ridership within the station area. In the future development, the planning and construction should be guided purposefully further to promote the integrated development of rail transit and the city.