Research On Distribution Features Of Home-work Location And Commuting Features Based On Multi-source Data In Shenzhen

The distribution of urban residents’ home-work place is the basis for research on issues such as urban transportation,spatial planning,and job-housing balance problem.In the era of big data,compared with time-consuming and laborious traditional questionnaires,the use of new data sources such as smart cards,mobile phone data,video cameras,coils,and social network data to obtain information on residents’ job and house place,and to analyze human mobility has become an international frontier research.Among these new data sources,mobile phone data and social network data have the characteristics of wide coverage,strong real-time data,and long sampling period,can record the spatio-temporal characteristics of user activities on a large-scale and long-term basis.However,because the two data have different characteristics,the existing research lacks a comparative study of the two,and there is no attempt to combine them for the study of residents’ job and house distribution and commuting characteristics.Based on the short-term rule sampling mobile phone data and long-term sampling of a large domestic social media data(Tencent instant messaging platform QQ user’s positioning data,hereinafter referred to as T platform social data),this article includes the following aspects:1)Feature analysis and processing of two data sources.2)Home-Work place identification.According to the characteristics of the mobile phone data and the T platform social data,two methods for identifying home-work place of residents are proposed.The method for estimating the coverage of the mobile base stations has been improved,and combined with the data of building types,a simple and effective mobile phone data identification method has been proposed.Logistic regression model was used to identify the job and house information in T platform social data.3)Analysis and verification of identification results.Combining data from census and residents,travel surveys,the results of the twanalyzed from a multi-scale and multi-angle perspective,and the sample bias analysis o data sources were compared and was also conducted.Finally,we summarizes the advantages and disadvantages of the two types of data in home-work research.4)Try to merge two data sources.Based on the characteristics of the mobile phone data and T platform social data,we try to combine the two data sources to improve the result,mainly from the perspective of improving the accuracy of mobile phone base stations to simulate to match resident’s home and work place.