Abstract:Urbanization in China has entered a new phase that emphasizes both scale expansion and quality improvement.This has led to demands on urban functional structures and rational urban planning.Street space serves as a vital spatial carrier for meeting urban residents'needs,such as travel,shopping,and leisure.It is comprised of urban roads and their ancillary facilities,buildings along the route,and many other elements.However,existing studies typically focus on the traffic function of urban roads,overlooking other functional aspects of street space as complex public activity areas,thereby hindering the optimization of street space quality.Therefore,there is a need to propose a classification method for the urban street space functions.Given the proliferation of taxi trajectory data and street view imagery,street space can be described in detail from the citizens'perspectives.Therefore,this study proposes a street space classification method that integrates taxi trajectory and street view imagery to delineate urban functions.The dynamic travel characteristics of urban residents in the street space are constructed using the taxi trajectory,including the number of trajectories passing by the street,the number of origin points on the street,and the number of destination points on the street.The physical environment characteristics of the street are constructed using the street view image,which contains single element street features,combined element street features,and overall element street features.Subsequently,based on residents'dynamic travel characteristics and the physical environmental characteristics of street space,the K-Means method is utilized to allocate street spaces with similar urban functions into the same clusters.Taking Bao'an District of Shenzhen as a case study,it was found that clustering with K=3 yielded the most interpretable results.Subsequently,based on street characteristics and auxiliary POI information(including POI density and POI enrichment index),street spaces were successfully classified into three urban functional types:commercial,traffic,and residential.Further analysis was conducted by integrating street view images,the Gaode map,and the 24-hours distribution of the trajectory,which validated the reliability of the classification results,achieving an experimental accuracy of 81%.Additionally,this study constructed two sets of ablation experiments for a comparative analysis of the effectiveness of the characteristics extracted in this study.The results of street space function classification based on taxi trajectory data and street view image data showed classification accuracies of 67%and 34%,respectively,which are lower than the classification effect of combining the two data sources.The results of street space classification based on feature selection from taxi trajectory data indicated that the classification accuracy considering only the weekday taxi trajectory features was 69%,and considering only the taxi trajectory features during T08:00-22:00 on both weekdays and weekends was 73%,which is lower than the classification effect of using three types of trajectory features throughout the entire day.By integrating the characteristics extracted from multiple data sources,the methods took into account both pedestrian and vehicle traffic functions,as well as the diversity of the streetscape environment and residents'activities.The identification results can provide references for the design and quality optimization of street space,as well as detailed investigations into urban functional areas such as residential,commercial,and traffic street space.

Discovering Urban Social Functional Regions Using Taxi Trajectories

Identifying Region-Wide Functions Using Urban Taxicab Trajectories

Recognizing Social Function of Regions with Taxi Traces

A multi‐modal transportation data‐driven approach to identify urban functional zones: An exploration based on Hangzhou City, China

Measuring social functions of city regions from large-scale taxi behaviors

Research on identifying urban regions of different functions based on POI data

Identifying Urban Functional Zones Using Public Bicycle Rental Records and Point-of-Interest Data

Identification of spatial and functional interactions in Beijing based on trajectory data

Identifying Urban Functional Regions: A Multi-Dimensional Framework Approach Integrating Metro Smart Card Data and Car-Hailing Data

Discovering Urban Functional Polycentricity: A Traffic Flow-Embedded and Topic Modeling-Based Methodology Framework

Urban Function Recognition at Street Level by Integrating Taxi Trajectory and Street-Level Imagery

Social Context Awareness from Taxi Traces

Urban Mobility Dynamics Based on Flexible Discrete Region Partition.

Social context awareness from taxi traces: mining how human mobility patterns are shaped by bags of POI.

Identification of Urban Functional Regions in Chengdu Based on Taxi Trajectory Time Series Data

Spatio-temporal Dynamics and Driving Mechanisms of Resident Trip in Small Cities

Discovering Urban Functional Zones Using Latent Activity Trajectories

Mining Spatiotemporal Patterns of Urban Dwellers from Taxi Trajectory Data

Discovering Functional Zones In A City Using Human Movements And Points Of Interest

Discovering Regions of Different Functions in a City Using Human Mobility and POIs

Integrating street view imagery and taxi trajectory for identifying urban function of street space