Street topological structure and cycling uses:A multiyear span data analysis based on gate count flow data in Tianjin
After the booming usage of shared bikes from 2017 in China,the number of cyclists has increased in many major cities.To promote non-motorized traffic became an important strategy for urban renewal and redevelopment under the goals of carbon peaking and carbon neutrality.Most research on non-motorized traffic borrowed the model for vehicles,focusing on the factors at city scale related to traffic demand and modal split.Other studies analyzed individual behavior choice at micro scale,focusing on the route choice or the spatial quality.However,there are very few empirical studies explored on the role of street pattern at neighborhood or street scale,dealing with the spatial distribution of non-motorized traffic.Especially,the emergence of shared bikes in China provides a valuable chance to study the changing tendency through a comparative analysis.This research is based on detail gate count data of 401 streets in 13 neighborhoods in Tianjin in 2014 and 2018.Using space syntax model,it studies the impact of street topological street pattern(measured by choice and integration values at different radius)at city-,neighborhood-and street scales.Other than topological street pattern,this research also analyzes factors such as distance to metro station,density(commercial POIs and building areas),street design(traffic light,railing,greenery,etc.)on the modal split and flow distribution of non-motorized vehicles.The findings are summarized as following.1)By comparing the amount and share of non-motorized traffic before and after the emergence of shared bike,data suggest that the increase of non-motorized traffic substitutes the pedestrian rather than vehicle traffic in all case areas.2)At neighborhood scale analysis on the share of cyclist usage,the density of shops(radii=200 m)shows strongest correlation on both 2014(R=0.88)and 2018(R=0.89).Topological street pattern(log choice r1km,R=0.61-0.63)and building density(Total building area,radii=200 m,R=0.60-0.67)also show clear correlation.However,these factors are strongly interrelated with each other that they cannot be all included in multi-variant regression model.All street design factors show very weak correlation,which indicate street design functions not at neighborhood scale.3)At street scale,this study group all gate count data into one model(401 gate count data)without average the data by the case areas.It also makes a distinction between the analysis on major streets as the border of neighborhood(101 gate count data)and minor streets inside the neighborhood only(300 gate count data).These three data sets also divided by two years(2014 and 2018).Six multi-variant regression model are established to analyze the flow intensity on each gate.The results show all data models and major streets in 2014 and 2018 can explain the data very well(adj.R2<0.6).Two minor street model have relatively lower R square value(adj.R2=0.47-0.55).Among all of these six models,topological street pattern has major impact on the non-motorized flow distribution.Building density factor(total building area in radii of 1 km)has minor impact.Additionally,there is a clear difference in the radius of space syntax parameters for major streets(7.5 km)and minor streets(2 km).These findings do not only demonstrate that the space syntax model could be used to analysis non-motorized traffic,but also further verify the linkage between individual behavior and the emergent pattern of movement for this group.The differences of scale factors in topological street pattern reflect the behavior pattern of short distance cycling by bike and longer distance movement by mopeds.As a summary,based on detailed gate count data in four years span,this research suggests"small block size"and"street continuity"are key features of street pattern to support non-motorized traffic,which may help to reach the goals of carbon peaking and carbon neutrality.
non-motorized transportationtypological street patternspace syntaxmodal splitgate count data