As China's urbanization process continues to advance and deepen, the spatial structure within each city is changing constantly. Analyzing the hidden relationship between human and land behind the massive urban data is a hot topic. In this paper, we apply the idea of network embedding to urban geography. The high-dimensional urban human activity data is embedded into the low-dimensional space (feature space) using the node2vec method. Compared to the traditional method of directly using high dimensional data analysis, node2vec can better detect the spatial relationship and structure of the network which are important for us to understand the urban functional areas more clearly. Based on Shanghai taxi data and subway travel data, we focused on the distribution and characteristics of functional areas in Shanghai. Then we verified the feasibility and superiority of node2vec.