Methods in urban temporal and spatial behavior research in the Big Data Era
QIN Xiao,ZHEN Feng,XIONG Lifang,ZHU Shoujia
DOI: https://doi.org/10.11820/dlkxjz.2013.09.005
2013-01-01
Progress in Geography
Abstract:The rapid development of information technology has taken us into the "Big Data Era",changed the organization and structure of urban space and residents' behavior,and also caused transformation of the methods in urban temporal and spatial behavior research.On the basis of summarizing the problems of traditional methods such as poor data accuracy,small sample size,weak continuity,and higher costs,this paper first combs through the data acquisition and processing technology for web data mining,residents' behavior data collection and analysis,and network map integration and visual development,which can affect the transformation of the research methods.Then it reviews the latest progress in applying big data to urban temporal and spatial behavior research at home and abroad from the perspectives of residents' behavior,urban space,and urban hierarchy,and builds up a method framework for urban temporal and spatial behavior research based on big data application.The methods in urban temporal and spatial behavior research are going through a great transformation because of the emergence of massive and various information data.Data collection methods have changed from yearbook statistics,social questionnaire survey,in-depth interview to mining of network data(social network data) and application of new spatial position technology(GPS,smart mobile phone,LBS,etc.),and the data shows obviously new characteristics such as large sample size,real-time dynamic,micro and detail,with more attention paid to the extraction of residents' geographic position information.However,as to specific research methods,the traditional ones are still widely used,such as descriptive statistical analysis,cluster analysis,factor analysis,gravity model,network analysis,space-time prism,etc.Generally speaking,the researches of urban temporal and spatial behavior have obvious characteristics of using "new" data and "old" methods to study "newer" and "older" problems at the present stage,and their research scope has also expanded from residential scale to urban space and regional range.However,problems still exist with the current research,such as how to eliminate fictitious data,how to learn and innovate analytical methods,how to expand research field and embody characteristics of the era.Therefore,it is necessary to promote the cross and integration of related disciplines such as sociology,economic geography,cultural geography,tourism geography,computer science,mathematics and geographic information science,in order to find new analysis methods,and also reinforce the research of residents' behavior and urban space by using social network(Twitter,Flikr,Facebook,Sina Microblog,etc.) data or other web(SouFun.com,Dianping.com,Zhaopin.com,Taobao.com,etc.) data,and guide innovation of urban planning methods.