Seeing Urban Transformations through Big Data: Evidence from China
Big data is increasingly regarded as a new approach for understanding human behaviour within and between cities. For example, the emergence of location-based social media (LBSM) data can potentially offer information with detailed human behaviour characteristics. The fundamental value of LBSM data is for public self-expression in real time—including when did it happen and where did it happen. There is a growing body of literature in exploiting these new sources of data, alongside traditional sources.
I am a Co-Investigator for the ESRC Urban Big Data Centre, and in this blog I set out some of our emerging findings. A primary objective of my work is to help in the construction of the big data evidence base for evaluating different aspects of urban transformations and related forms of geographical division of social-spatial mobilities and city liveliness. China has experienced rapid, albeit geographically uneven, economic growth during the past few decades. Notwithstanding the uneven development pattern, the government has gradually relaxed the institutional constraints on human mobility between the more affluent coastal cities and less developed inland regions. However, the existing research has mostly overlooked the spatial-social-cultural aspects of human behaviour at fine spatial and temporal scales. We believe that this oversight results from the difficulty of identifying individuals’ mobility footprints, calculated from location-based big data such as from mobile phones or WeChat and Weibo social media platforms.
At the inter-city level, our recent study provides new evidence on the relationship between cultural ties and human mobility at the aggregated city level in China. Today, China’s social media market is booming, and this rapid expansion means that people can easily access the internet for posting their footprints. From this we can extract geo-tagged information to show millions of individuals’ mobility footprints by cities in mainland China (see photo at the top of this blog). Figure 2 further illustrates the distribution of human mobility flows between core cities (Beijing, Tianjin) and peripheral cities like Langfang, Baoding in the Great Beijing Metropolitan Area. The main finding is the significant effects of cultural diversity on human mobility between city pairs. By exploiting sub-sample characteristics, additional results provide the insights that such effects are not distributed evenly by travel motivations and between core and peripheral regions. Future studies should seek to understand the more precise mechanisms that facilitate human mobility and long-term migration across cities.
At the intra-city level, our recent work studied the complexity in the geographical connection between city liveliness and spatial configurations for consumptive amenities. Our study uses a spatially-temporally varying estimation approach to present a step towards understanding this question though location-based big data perspectives. The common perception of “city liveliness” is that it corresponds to the rise and decline of hectic city living patterns, with the variation of time-space aggregate human activity. With the emergence of big data, human activity patterns are increasingly tied to social media platforms via mobile phones for productive business, social interactions and leisure. A small and growing body of literature uses big data visualization techniques to map and model human activity patterns. In our study, aggregated space-time human activity intensities are measured by using mobile phone positioning data on an hourly basis (Figure 3). The mobile phone positioning data are obtained from the location-based social network service (SNS) provider Tencent in China. Consumptive amenities such as restaurants, shopping malls are identified by point-of-interest data from the Chinese Yelp website (Figure 4). Using Beijing as a case study, we find that consumptive amenities affect dense and diverse human activity patterns. Additional spatial modelling results provide the insights into the geographic contextual uncertainties of consumptive amenities in shaping city liveliness. Future studies are encouraged to integrating the importance of the co-location of different types of consumptive and productive amenities in driving the city liveliness and productivity.
These urban big data studies have implications for geographers and planners seeking to understand the geography of urban transformations in developing countries where traditional census and surveys are relatively poor. I acknowledge that there are potential limitations of urban big data applications such as sampling representativeness and limited information about individual socioeconomic backgrounds. However, I expect a bright future for the use of big data in urban analysis. Our recent studies offer some useful ‘food for thought’ for future work and encourage more efforts in improving the quality of empirical assessments to contribute to our understandings of urban transformations through big data.