As the most important parts of urban systems, jobs and housing spaces and their balance directly affect the spatial structure of cities, the behavior and experience of the residents, and the harmony and livability of the society. This study used more than 100 million records of mobile phone signaling data, covering the whole city of Beijing and over a period of one month, to identify the jobs-housing spaces by targeting the origin-destination (OD) oriented connections applying the density-based spatial clustering of applications with noise (DBSCAN) method. Furthermore, this study explored the spatial distribution pattern and matching characteristics of Beijing's jobs and housing spaces from different spatial scales by using various measurement methods of jobs-housing balance. The spatial scales for analysis cover the whole city, ring roads, districts, and residential community and town and townships, and the methods applied include the spatial mismatch index, deviation degree, commuting flow rate, and so on. The results show that: 1) Mainly influenced by the spatial layout of the Beijing Master Plan, Beijing's housing space is characterized by dispersion at the large scale and agglomeration at the small scale, showing a pattern of scattered groupings. In contrast, its working space presents features of agglomeration at the large scale and dispersion at the small scale, retaining a significant single-centered layout. 2) Although working in local areas is the first choice for people at both the ring road scale and the district scale, there are still a great number of people works outside their residential areas. The degree of jobs-housing mismatch gradually decreases from the central city to the periphery no matter which method was adopted or at which scale. At the residential community and town and township scale, however, a more detailed feature of three-zones, with job agglomeration inside, housing agglomeration in between, and balanced distribution outside, was observed. 3) Both the general lack of jobs in certain areas caused by the high concentration of working space and the two-way commuting phenomenon in the majority of the areas caused by the high spatial concentration of jobs indicate the necessity of spatial reorganization of residential function and employment function. Specifically, for the regions with a high proportion of two-way commuting flow between them, such as Chaoyang-Changping, Tongzhou-Daxing, Haidian-Changping, Mentougou-Shijingshan, and Chaoyang-Tongzhou, further in-depth investigation should be conducted to find out the reasons for its formation and then possible industrial adjustment or functional reconstruction from the city level should be coordinated. Relying on big data, on the one hand, the job preference and demand of local residents can be identified, so the types and number of jobs in each region could be adjusted accordingly; on the other hand, the proper locations for new job and residential centers may be identified to help rearrange the land use of the whole city.