Abstract
Abstract
The large human migration between different areas in China indicates unequal urban and economic development under the background of rapid urbanization and rush for economic growth. Based on social networking big data, we analyzed the difference of population distribution during the Spring Festival holiday and the non-holiday periods. We find that the two patterns differ significantly and there are mainly two levels of population migration center.
The Chinese Spring Festival travel season is a period of travel in China with extremely high traffic around the time of the Chinese Spring Festival, which is the largest annual human migration in the world (McCarthy, 2018). It has occurred since the Chinese economic reforms of the late 1970s, after large numbers of rural laborers flocked to the cities to find jobs. It is reported that around 2.97 billion passengers traveled during the 2018 Spring Festival travel season from February 1, 2018 to March 12, 2018 (Feng, 2018). Among them, most people were recorded twice or more due to the round-trip and other trips they made. A high level of human migration between rural and urban areas not only indicates economic inequality and urban development but also brings problems of left-behind children, educational inequality, land-use change, environmental change, and more (Wang et al., 2019; Wei et al., 2018; Xu et al., 2017).
We analyzed two typical periods of population distribution based on social networking big data from the Chinese tech giant Tencent. Tencent offers rich, internet-based products and services. For example, QQ and WeChat are two of their multipurpose instant-messaging mobile apps, each of which has approximate 900 million active users. Based on location query services, Tencent can record a user’s geographical location. Due to its large number of active users, we argue that user distribution could be used to reflect how people are distributed and have migrated in the country, although there might be some bias similar to that seen in other social networking data (Graham et al., 2013; Liu and Wang, 2015).
We collected the daily heatmap data published by Tencent during the Spring Festival holiday from February 16 to February 20, 2018, and during the week after the Spring Festival travel season from March 12 to March 16, 2018. The data use 0.01 × 0.01 degree resolution. The grid value of the data records the number of visitors per day. The daily average total number of active users during the Spring Festival holiday and the non-holiday period was approximately 799 million and 998 million, respectively.
Users in prefectural-level cities during the two periods were aggregated from the two average grid datasets and normalized by the corresponding sum of all grid values (Figure 1). The maps were re-projected by replacing the geographical area with population (percentage of total) (Figure 1(a) and (b)), or with ratio values of the difference between the population (percentage of total) during the usual non-holiday and the population (percentage of total) during the Spring Festival holiday (Figure 1(c)). The Pearson correlation coefficient between the population during the Spring Festival period (Figure 1(a)) and the population from the sixth national census in 2010 is 0.93, which is higher than the coefficient between the population during the usual non-holiday period (Figure 1(b)) and the population from the sixth national census: 0.89. The substantive correlations suggest that the Tencent data could well reflect the urban population size. It might also indicate that because of rapid urbanization and the high level of human migration across the country, the spatial pattern of the usual de facto population distribution has significantly changed since 2010. Several other conclusions could be drawn from the figures.

Cartogram of prefectural city population during (a) the 2018 Spring Festival (SF) holiday period and (b) the usual non-holiday period; (c) the population change ratio between (b) and (a).
First, Chongqing is the city with the largest population during the Spring Festival holiday, accounting for approximately 2.21% of China’s total population, which is about twice the population of the second-largest city. Besides Chongqing, there are four cities with populations representing more than 1.00% of the country’s total: Baoding, Chengdu, Beijing, and Handan. However, Shanghai, the important economic center of China, accounts for only 0.85% of the population during the Spring Festival holiday.
Second, during the non-holiday period, the top four areas of population accumulation can be clearly seen in Figure 1(b). They are the Yangtze River Delta region (Shanghai, Suzhou, Hangzhou, Nanjing, etc.), the Pearl River Delta region (Guangzhou, Shenzhen, Dongguan, Foshan, etc.), Beijing–Tianjin–Shijiazhuang, and Chengdu. In addition, several typically smaller cities attracted large populations (Figure 1(c)), including Xiamen, Kunming, Wuhan, Changsha, Guiyang, Zhengzhou, and Xi’an, most of which are developed regions or provincial capitals.
Third, this data could be used to estimate fine spatial and temporal scale population distribution and be used in further research, such as seasonal migration, air pollution exposure, and energy consumption.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (grant number XDA23000000), the National Science and Technology Major Project (grant number 2017ZX10201302), the National Natural Science Foundation of China (grant number 41771434), and Innovation Project of the State Key Laboratory of Resources and Environmental Information System (grant number O88RA200YA).
