Yao Yang,Zhou Xiaojin

(Guangzhou Academy of Social Sciences,Guangdong Guangzhou 510410)

?

Research on Population Flow of Megacity Based on Big Data of Megacity〔*〕

Yao Yang,Zhou Xiaojin

(Guangzhou Academy of Social Sciences,Guangdong Guangzhou 510410)

The traditional census or population survey data is increasingly turning into big-datalization.The analysis based on mobile communication,and social network data is more helpful to the national or regional population inflow,outflow and its changing rate research,and becomes more accurate,dynamic and real-time.In the background of accelerating the implementation of the National Big-data strategy,we should speed up the sharing of public data,while the purchase of data for big-data research is even more important and urgent.

population flow;big-data of population;national big-data strategy

Ⅰ.Research progress on population flow based on big data analysis

Domestic and foreign scholars have been carried out more research and exploration,mainly focus on mining and analysis the data from twitter(Naaman M,Zhang A X,Brody S,et al.，2012),Facebook,Sina Weibo(Zhen Feng,Wang Bo,Chen Yingxue,2012) and other social network data (the location of the user),mobile phone call data(Krings G,Calabrese F,Ratti C,et al.，2009；Kang C,Zhang Y,Ma X,et al.，2013),intelligent transport card data and urban monitoring data,through the study of urban network activity to determine intercity relations and hierarchical system,simulating urban residents travel mode and the movement of population change to guide more reasonable transportation planning and land use adjustment.

In 2014,domestic related research literature about population big data is increasing,but they mainly focus on method breakthrough.The empirical studies are inadequate;the floating population research based on big data of Guangzhou is in a blank state.Hu Qiaoling,Ru Jinping (“Computer Simulation”,2014) use the improved algorithm for big data statistics of the population migration,to improve the accuracy of the prediction of population migration.Through the data analysis and data mining,Wang Feng and Tang Meihua (“Mobile Communication”,2014) analyze the urban population spatial and temporal distribution and dynamic migration.Zhao Shiliang,Gao Yang (“the Population and Society”,2014) point out that using the broadcasting mechanism between mobile phones and mobile communication base station,it can be analyzed such as residential vacancy rate and urban population commuting scale and flow.Li Hongjuan (“Modern Management Science”,2014) conducted exploratory study of China’s population information management and application in the era of big data.Liu Yu,Kang Zhaogui and Wang Fahui (Journal of Wuhan University,2014) discussed the method of constructing the model of the observed movement pattern.Tong Dahuan (Tencent blog,2014) for the first time using QQ big data,analyze population mobility in the first-tier cities:Beijing,Shanghai,Guangzhou and Shenzhen.Zhang Qiang,Zhou Xiaojin (2014) using the total amount of mobile communications data estimated country’s major urban population,the results are quite consistent with the latest city surveys of the population by Beijing,Shanghai and other cities.Different from the traditional population data relying on survey or census,the research of population flow based on big data is carried out by professionals in the field of information technology.While the research achievements in the field of traditional demographic will have contributed to the effectiveness of analysis and judgment of data classification,clustering,regression and correlation.

Ⅱ.Population flow analysis of Shanghai based on QQ user login

Big data in social networks,QQ users login to the Tencent Inc and the real-time analysis of Wechat users can obtain the population distribution and flow data accurately.Because the QQ user age mainly between 18 to 50 years old,the age is also consistent with the age structure of the floating population.Thus,by analyzing the changes in mass QQ login area during the Spring Festival,the city area population movements in this age can be calculated.Tong Dahuan (we Tencent blog,2014) for the first time using QQ big data analyzed of the population flow in first-tier cities of Beijing,Shanghai,Guangzhou and Shenzhen.Tongda Huan believes,containing instant migrants,the number of actual population at the end of 2013 in Beijing,Shanghai,Guangzhou and Shenzhen four cities is not the official 69.3 million,but up to 164.76 million.Tong’s error is simply take QQ users equated with the population,but ignoring such a crucial fact: active QQ users can log onto the computer,on the other hand more through mobile phones and other mobile terminal to log in,mobile phones per capita in this age group in Beijing,Shanghai,Guangzhou and Shenzhen four cities is 1.5.taking these key factors into consideration,the population within age of 18-50 years old in first-tier cities of Beijing,Shanghai,Guangzhou and Shenzhen is estimated at 64.14 million,plus the household population other than the population of this age in the four cities,as well as the inflow of floating population other than the population of this age,would be the total population of the four cities.In addition to the inevitable considering the population structure,the use of QQ login when urban population estimates must take into account the frequent business travelers who log on in certain regions phenomenon.

Collected QQ user data set are consisted of five fields: the serial number,grade,active degree,provinces and cities.The latest acquisition of the four major cities of QQ user data set contains 180 million data,filtered the lower activity travel and short-term users whose activity are less than or equal to 30,the results is showed in Table 1,column one data.QQ user login corresponds to non-short-term population forms the second column of data;Every 1.4852 active QQ number corresponds to a post 80s or post 90s (after calculation,the post-80s and post-90s per capita 0.85 QQ number,while the four major cities in the age of the population has the QQ number per capita rate of nearly 1.5.The fourth column number represents the post 80s and post 90s population in each city.Then,Use the proportion of this age population accounts for urban population (column 5),we can infer the total urban population (column 6).

Table 1 The estimated population (2013) in China’s first-tier cities based on QQ user login data

Using QQ user login address changes,during 2014 Spring Festival (January 16-February 1),we Calculate the total of 33.412 million young people left Beijing,Shanghai,Guangzhou,Shenzhen (QQ user groups).Beijing,Shanghai,Guangzhou,Shenzhen outflow population before the Spring Festival respectively is 12.688 million,10.433 million,10.493 million and 10.012 million.The population stay in the cities during Spring Festival respectively is 10.563 million,14.658 million,9.587 million and 8.53 million,a total of 42.86million in Beijing,Shanghai,Guangzhou,and Shenzhen for Spring Festival,while the outflow of the total population is up to 43.63 million (Table 2).Tong Dahuan (Tencent we blog,2014)believes that the 2014 Spring Festival,Beijing,Shanghai,Guangzhou,Shenzhen were suffering a 10.700 million people (users) permanent escape,but that is not the case.Because four cities are famous tourism cities in China,QQ login address of the corresponding change is also reasonable,it is still unable to determine exactly how many people leave four cities if we only estimate through the QQ login address change.Similarly,after the Spring Festival a large number of new QQ users login four cities,does not mean that a large number of people rush to the big cities in search of work,it may due to more tourists visit the four cities during the Spring Festival.For example,during the seven days of 2015 Spring Festival holiday,tourists from other cities to Beijing reached 1.425 million,up 7.1% compared with the same period in 2014.

Table 2 The estimation of urban population movement in China’s first-tier cities based on QQ user login data (2014)

Ⅲ.Population flow analysis based on big data from geographical distribution of internet users

Compared with QQ user that concentrated in the age of 15-35,the mobile network access represented by a mobile phone is more extensive.The population over 15 years all has a mobile phone.Early in 2007,Guangzhou migrant workers had a mobile phone ownership rate of more than 100%,mobile phone ownership rate of non local population was nearly 100%,which is higher than local household population.With the popularity of mobile phone,and the rapid increase of mobile phone penetration rate of Internet users,the geographical distribution has become effective tools and channels of population flow analysis.

The Spring Festival population migration is a rare opportunity for population flow research.According to Nanfang Daily (Reporter /Cheng Xi Communicator/ Transportation propaganda) reported that during 40 days (February 4-March 15) in 2015 Spring Festival,Guangzhou sent a total of 29.5364million passengers,an increase of 3.06%.Among them,railway send 12.1759 million passengers,an increase of 16.07%;Highway send 13.9512million passengers,down 5.79%;Waterway send 41,000 passengers,down 16.50%;Civil Aviation sent 3.3683 million passengers,an increase of 5.62%.With the operation of Guiyang-Guangzhou,Nanning-Guangzhou Shanghai-Kunming high-speed railway,the railway transport capacity increase is large,especially the high-speed rail passenger traffic growth,During the Spring Festival,high-speed rail passenger sent by Guangzhou South Railway Station has exceeded the Guangzhou Railway Station and East Railway Station,about accounting for 40% of total railway passenger flow.Before the peak transport in the Spring Festival,the returning home force has alreadly started,based on the flow of students and self-employed passengers.With the plants holidays began,migrant workers became the main force of passengers returning home,while passenger flow of rural retaining population came to Guangzhou for Spring Festival gradually increased.During the Spring Festival,tourists are the main force of passenger flow.After 7th Jan of lunar year,the retaining population returned back to home.From 7th Jan to17 th Jan of the lunar year,it is the peak of return journey of floating population back to Guangzhou.

In 2015 during the Spring Festival,Guangzhou Internet users accounted for the proportion of 4.24% on 2015 January 31 (lunar year December 12) and decreased to 4.34% of 19 February 2015 (the first day of the Spring Festival).As of the end of December 2014,Internet users in the country is 649 million,which projected the net outflow of population out of the province during the Spring Festival in Guangzhou is 584.1 million,On Jan 2nd of the lunar calendar,further decline in the proportion to 4.29%,it is estimated that the local population out of the province for travel during the Spring Festival reached 324 500 people.According to the National Planning Commission,Department of floating population in our earlier data projections,the floating population from Guangdong province in Guangzhou,accounted for 66%.During the Spring Festival in Guangzhou,population estimated net outflow is 8.85 million,compared with the previous estimation of migrant population based on QQ login;both only have a difference of 50,000 people.Take 8.8 million population estimated from QQ login data as the basis,the calculating error is quite small (0.57%).Using the same method,the projected net increase of 2015 Spring Canton Fair is 794800 passengers.

According to the Baidu 2015 spring migration data,for 2015 Spring Festival holiday,Beijing,Shanghai,Guangzhou,Shenzhen,Dongguan outflow population accounts for the country’s proportion were 8.27% 、6.11%、4.44%、6.91% and 5.22%,the inflow population accounts for the country’s proportion were 2.77%,1.93%,0.4% and 0.3% and 0.2% respectively,five city net inflow of population accounts for the country’s proportion were 5.50% and 4.88%,4.04%,6.61% and 5.02%,respectively.Guangzhou’s net return home population outside Guangdong province is 6.06million,which is the 5th in China,ranks after Shenzhen,Beijing,Dongguan and Shanghai.The population of spending Spring Festival in Guangzhou is 600,000.According to Baidu Spring Festival migration big data,the estimated population flows in other cities as shown in Table 3.

Table 3 Baidu 2015 spring festival population migration big data inference

Ⅳ.Conclusions and policy suggestions

The 13th five year plan put forward the concept of co-construction and sharing of social governanc,which is to change the previous governance mode led and promoted unilaterally by government,to encourage the different interests subjects and action subjects to participate in social governance.Herefore,social governance and public services will meet the needs of all kinds of social groups,in order to achieve a harmonious situation of governance.Population big data as the basis of modern social governance also requires enterprises,research institutions and governments to build,to share and to participate in.Government data in the process of getting the process often need to pay a huge cost.The cost of traditional process of getting data for governmnet is very high.Taking the census as an example,a census every five years will directly cost 50 billion yuan paid by taxpayers,with an average annual direct cost of more than 10 billion yuan.This kind of human based survey will inevitably have a wide range of errors.While the population flow analysis based on big data from mobile communication,traffic and social network based will help to provide a more precise and accurate population basic information.What’s more,these basic information can be dynamic,real-time updates and visual presentation.

In order to promote the research on megacity population flow based on big data,and to accelerate the application in modern social governance of megacity.The authors suggests as follows:

1.Construction of big data from different sources,each with a focus and complementary

This paper summarizes the traditional census,population survey,and other big data analysis methods based on mobile communications,public transportation,social networks,and urban water utility.The above methods are divided into two categories: the 1st category includes the population digital information and the data with different forms,which is continually collected with a clear purpose,carried out according to plan,amd led by government departments.It can be called as structured big data or ordered big data;the 2nd category includes a large number of morphological data generated by random generation,due to the rapid development of new technology based on Internet and Internet of things,which can be called as unstructured big data or random big data.And from the perspective of the basis of social governance and the demand of fine management,these two categories of big data and the sub small class of diverse data have great application value.Only when these complementary data from various sources,and each with a focus can be used in comprehensive mining and appliacation,they will play the biggest effect of big data on the governance of government.Minewhile,the Internet,Internet of things,cloud computing and other technologies also provide technical support for the intelligent level of data collection.

2.Innovation of megacity social governance mode

With the megacity’features of huge number of population,population mobility and high density of population,the city sickness is increasingly worsening,such as increasing heterogeneity of social structure,serious differentiation of social groups and differentiation between the rich and the poor,diversified cultural concept and interest demands,traffic congestion,insufficiency of public transport capacity,limited education resources,shortage of medical services.All these social characteristics of megacity are a series of challenges to the implementation of fine social governance.Thus,the government should actively innovate the thinking mode of megacity social governance.The arrival of the era of big data brings to the government with opportunities and challenges coexist.Big data is not only a technical change,but also the innovation of government management thinking and mode.The government should actively become an important developer and user of the population big data,to play a full role in the development of big data market resources,mining big data implicit correlation,embodying the application value of big data.Thus,the government is able to provide more complete and perfect service to the society through the accurate,dynamic and real-time data,and build a smart government,and realize the development of smart city.

3.Environment for development of big data research and applications

The information generation of the national population base is still based on enterprise data,especially for the analysis of large data such as mobile communications,transportation and social networks.Whether it is from point of view of the cultivation of big data professional personnel or to promote the modernization of national governance,mobile communication,transportation,and social network operators should provide the data interface after the private information has been filtered to universities and research institutions for their big data research as far as possible.We suggest that in order to promote the national population data research,the government should clear legislative requirements to mobile communication operators,public transport operators (such as subway,bus,high-speed rail,aviation enterprises,etc.),BAT and other enterprises that have accessibility of big data to provide for the interface of data or data for research.Enterprises that provide big data for reasearch also provide the necessary data processing,which will be paid by government (government procurement).The service of the data provider can also be compensated in the form of tax deduction.As for universities or research institutions,they should also clear data needs.Country can set up big data base research data fund.And national basic research data should be open to universities and research institutions.

Notes:

〔1〕Balázs Cs.Csáji,Arnaud Browet,V.A.Traag,Jean-Charles Delvenne,Etienne Huens,Paul Van Dooren,Zbigniew Smoreda,Vincent D.Blondel.Exploring the mobility of mobile phone users,Physical A: Statistical Mechanics and its Applications,2012,pp.1459-1473.

〔2〕Gonzalea MC,Hidalgo CA,Barabasi AL,Understanding individual human mobility parrerns,Nature,2008,453(7196),pp.779-782.

〔3〕Song C,QU Z,Blumm N,et al.,Limits of predictability in human mobility,Science,2010,327(5968),pp.1018-1021.

〔4〕Bao Ting,Zhang Zhigang,Jin Cheqing,An urban population flow analysis system based on mobile big data,Journal of East China Normal University(Natural Science),2015(05).

〔5〕Chen Jia,Hu Bo,Zuo Xiaoqing,Yue Yang,Driving Behavior Analysis Based on Trajectory Data Collected with Vehicle-mounted GPS Receivers,Geomatics and Information Science of Wuhan University,2014(06),pp.734-738.

〔6〕Hu Qiaoling,Ru Jinping,Simulation on the Prediction Model for volume of Population Migration Based on big Data Analysis,Computer Simulation,2014,31(10),pp.246-249.

〔7〕Liu Yu,Kang Chaogui,Wang Fahui,Towards Big Data-Driven Human Mobility Patterns and Models,Geomatics and Information Science of Wuhan University,2014(06),pp.660-666.

〔8〕Wang Feng,Tang Meihua,Urban population management solution based on big data of mobile communication,Mobile Communications,2014,(13),pp.38-41.

〔9〕Wang Guangzhou,Research and Innovation in the Population Science of China in the Era of Big Data,Population Research,2015(5).

〔10〕Zhang Qiang,Zhou Xiaojin,Population Size Estimation and Regulation Path for China’s Big Cities,West Forum,2014(2),pp.1-16.

About the author：Yao Yang(1979—),associate researcher of Economics,MPA of California State University,research direction:regional development and Local governance，Urban Economy;Zhou Xiaojin(1971—),researcher of Economics,Ph.D.in Economics/ Postdoctoral of Finance,research direction: Urban (Population)Economy,High-speed rail economy,Big Data Applications.

〔*〕General project of National Social Science Fund 2015:Research on Population Migration,Population Flow and New Change of Directions Based on Big Data(15BRK037).This thesis is the staged achievements of Guangzhou humanities and social science research base“Guangzhou N ational Central C ity Research Base”research project.