基于手机数据可靠性分析的旅游城市功能空间识别研究

HOW RELIABLE ARE CELLULAR POSITIONING DATA IN TOURISM ENVIRONMENTS?AN EXPLORATION OF FUNCTIONAL REGIONS

随着移动通信和移动互联网的普及,带有地理空间信息的"大数据"受到人文地理学者的关注。本文关注手机移动数据的旅游研究价值,以张家界市武陵源区、慈利县为例,首先分析了通信公司漫游用户数、基站网流量和话务量代表实际游客量的可靠性程度,然后探讨了漫游用户数波动对大尺度行政区的功能空间反映,最后挖掘了网流量波动对小尺度基站区的功能空间识别。结果显示,漫游用户数仅在旅游区对实际游客量具有可靠性,基站网流量比话务量的可靠性更强;漫游用户数在旅游区的波动幅度更大;基站网流量的空间因子分析识别出日常居住区、重点旅游区等功能空间。

Following the wide use of mobile communications and mobile Internet, human geography scholars have paid considerable attention to big data with geospatial information. This research focuses on the value and exploration of cellular positioning data in tourism research and has detected functional regions in a tour- ism-oriented city. This empirical case study is based on Wulingyuan district and Cili county in Zhangjiajie city. The dataset contains not only the aggregated records at the cell tower level from China Mobile and at the administrative area level from China Unicom, but also the real tourist numbers from government offices. The empirical study first explores the reliability of roaming user numbers, and Internet and call traffic vol- umes. The results show that roaming user numbers are only reliable in tourism areas and Internet traffic is better than call traffic if comparing their correlation levels with actual tourist numbers. This finding might be associated with personal cellular phone use habits. The study then analyzes the temporal rhythm of roaming user numbers to find the characteristics and differences between tourism and residential areas. The results show significant differences; i.e., roaming user numbers increase sharply around holidays in tourism areas while the changes are relatively slow in residential areas. According to the reliability discussion on Intemet traffic volume, the final part of the study applies an exploratory factor analysis to the seasonal tourism spac- es in Zhangjiajie city based on these data sources, which could allow deep explorations at small spatial scales. The functional regions discovered by cellular positioning data are mostly correlated with physical space characteristics, which shows that the exploration is probably valid and reliable.