Research On Interaction Between Urban Public Transportation And Land Use Using Big Data

In the process of rapid urbanization,urban transportation problems have become a common phenomenon in the development of metropolis,which has reduced the city's operating efficiency and the level of public transport services.Although major cities have upgraded the public transportation infrastructure and constructed rail transit system,due to the lack of quantitative research on the interactions between urban public transportation and land use,the traditional transportation planning and management methods are not effective in solving urban transportation problems.Therefore,understanding the interactions between urban public transportation and land use is an important prerequisite for formulating scientific and reasonable urban planning.In the era of big data,the emergence of smart card data has brought new data and new methods to quantitatively study the interactions between urban public transportation and land use.The smart card data contains fine-grained individual travel information,which accurately reflects the spatio-temporal characteristics of passengers' travel flow and travel demands,which can further improve the validity and scientificity of the results.Based on big transportation data,this study explored the interactions between urban public transportation and land use,and the main findings were shown as follows:(1)Urban multi-modal public transportation is a complex system consisting of stations,lines,and passengers' travel flows.Previous studies had constructed public transport network models based on a single transportation network or clustering multiple transport networks due to the lack of the passengers' travel data.These simplified public transport network models were difficult to represent the true topological characteristics of urban public transportation and passenger capacity.Based on big transportation data,this study expanded the single-layer public transport network model by assigning weights of travel demand and efficiency to construct the bus-metro multilayer network.The analysis of the topological properties in the bus-metro multilayer network showed that compared by single-layer public transport network,the bus-metro multilayer network had more accurate in reflecting the topological structure of public transport network and the spatial distribution of passengers' travel flows.In order to further verify the applicability,effectiveness,and practical significance of the bus-metro multilayer network,this study used the urban scaling law to explore the scaling law characteristics of the bus-metro multilayer network at different spatial scales.The scaling relations between population distribution and node's weighted degree of the bus-metro multilayer network can be applied to reveal the balance between public transport service and population size,which will benefit the planning of urban public transportation and land use.(2)The coordinated development of urban public transportation and land use is the prerequisite for solving urban transportation problems,and the accessibility of public transport network is the key to understanding the interactions between urban public transportation and land use and grasping their discipline.From the perspective of public transport network accessibility,which generally exhibits spatial heterogeneity in urban regions,using POI(Point of Interest)data to evaluate the spatial distribution characteristics of land use can help refine urban land use management.Based on the bus-metro multilayer network,this study constructed the direct and transfer networks and proposed the calculation method for the accessibility of the bus-metro multilayer network.Using the accessibility of each grid as the dependent variable and the six types of POIs as explanatory variables,the geographically weighted regression(GWR)model was used to quantitatively analyze and evaluate the spatial heterogeneity of urban land use distribution.Therefore,based on the integrated analysis of POI data and GWR model,the public transport network accessibility could be used to evaluate urban land use and enriched existing evaluation methods for the study of interactions between urban public transportation and land use.(3)Passengers' travel flow is an important reflection of the non-linear impact of land use on urban public transportation and is also an important goal of land use planning under the guidance of public transportation oriented development,which represents the coupling relationships between urban public transportation and land use.Previous studies have demonstrated that the linear model cannot accurately express the non-linear characteristics of the influence of factors(such as land use)on passengers' travel flow of the metro stations.This study divided the metro stations into five types based on passengers' travel flow and POI data: work station,residence station,public facility station,work-mixed station,residence-mixed station.On this basis,XGBoost(e Xtreme Gradient Boosting)model was used to study the non-linear characteristics of the impact factors(station type variables,land use variables,and traffic characteristics variables)on passengers' travel flow of the metro stations.In order to solve the black box characteristics of the non-linear machine learning model(The inexplicability of the influencing factors is insufficient),this study introduced the SHAP(SHapley Additive ex Planations)model to quantitatively analysis of the non-linear characteristics of factors in the XGBoost model: importance,direction,and threshold phenomena.Interpreting the non-linear characteristics of the impact factors on the passengers' travel flow of the metro stations has a strong reference and guiding significance for most cities that are developing the rail transit construction.