天气因素对武汉市出租车出行活动的影响
作者简介:康朝贵(1986-),男,湖南衡阳人,博士,副研究员,研究方向为时空大数据分析与挖掘。E-mail:cgkang@whu.edu.cn
收稿日期: 2018-03-02
要求修回日期: 2018-11-24
网络出版日期: 2019-01-20
基金资助
国家自然科学基金项目(41401484、41830645)
国家重点研发计划项目(2017YFB0503604)
Impact of Weather Condition on Intra-Urban Travel Behavior: Evidence from Taxi Trajectory Data
Received date: 2018-03-02
Request revised date: 2018-11-24
Online published: 2019-01-20
Supported by
National Natural Science Foundation of China, No.41401484, 41830645
National Key Research and Development Project of China, No.2017YFB0503604
Copyright
天气状况作为人们生活环境的组成要素之一,对居民日常出行可产生显著的影响,具体可表征为特定空间位置和用地类型范围内出行活动的需求量以及道路交通路线选择的变化。高效、智能化的交通应急管理和城市规划建设亟需理解天气因素影响交通出行时空分布的基本规律。本文选取武汉市作为典型研究区域,基于出租车、气象和空气质量等数据,对不同天气下的居民出行模式和司机路径选择模式进行时空分析,并解释2类模式产生变化的原因和机制。结果表明:① 从时间上看,工作日的出租车需求量更容易受到天气变化的影响,其中降雨、气温的升高和风速的增强会显著降低居民对出租车的需求;② 从全市域空间尺度上看,降雨使得居民对出租车的需求量在工作日时段减少,而在周末时段增加,其中降雨主要刺激短距离出租车出行需求而抑制中长距离出行需求;③ 从城郊区空间尺度上看,雨天时段主城区内部的中距离流量减少,郊区内部的短距离流量增加,往返于主城区和郊区的中长距离流量在工作日减少、在周末增加;④ 从功能区空间尺度上看,下雨使得行政办公用地的出租车需求量减少,商业金融用地的出租车需求量在工作日减少、在周末增加,工业用地的出租车需求量在工作日增加、在周末减少;⑤ 从行驶路径上看,出租车司机在晴天时偏好根据距离来判断最佳路线,而在雨天倾向于改变原先路线选择策略,将距离和车速共同作为最佳路线的指标,选择用时最少的最佳路线。本文研究成果可帮助城市和交通管理部门更加深入地理解城市居民出行规律及其时空分布特征。
康朝贵 , 刘璇 , 许欣悦 , 秦昆 . 天气因素对武汉市出租车出行活动的影响[J]. 地球信息科学学报, 2019 , 21(1) : 118 -127 . DOI: 10.12082/dqxxkx.2019.180122
Weather conditions have a substantial impact on urban residents' daily travel activities. They usually determine the travel demand within a specific spatial location by land use type, as well as the route selection strategy between a pair of travel origin and destination. This information is crucial for stakeholders including urban dwellers, city planners and transport managers to optimize urban mobility, facility allocation and transportation resilience. In this paper, we apply spatiotemporal statistics, multiple linear regression and clustering analysis on taxi data and weather records of Wuhan City, China to understand the spatiotemporal characteristics of residents' travel demand and taxi drivers' route selection under different weather conditions. As a result, the dominant weather condition factors influencing residents' travel activities are revealed on space and time. First, taxi demand is more vulnerable to weather changes on weekdays than weekends. It is negatively proportional to the increasement of rainfall, temperature and wind speed. Second, at city scale taxi demand decreases along with raining on weekdays while the demand increases on weekends. In particular, the short-distance travels increase while medium- and long-distance travels decrease. Third, taxi demand is more vulnerable to weather changes within the urban area than the suburban area. On rainy days, medium-distance travels within the urban area decrease, whereas short-distance travels within the suburban area increase. Fourth, taxi demand on residential area increases, whereas the demand on commercial area decreases on rainy days. Last, taxi drivers are found to prefer the shortest path on sunny days and the fastest path on rainy days. Those research results can assist urban planners and municipal managers to enhance their understanding of urban residents' mobility pattern and their spatiotemporal dynamics more deeply.
Tab. 1 Regression model for taxi demand and weather condition on city scale表1 出租车需求变化与天气因素全局回归分析结果 |
变量 | 工作日 | 周末 | ||||
---|---|---|---|---|---|---|
原始回归系数 | 标准化回归系数 | 原始回归系数 | 标准化回归系数 | |||
Intercept | 62.041*** | 16.324 | ||||
-190.120*** | -0.281*** | 31.798 | 0.029 | |||
-137.088*** | -0.132*** | 34.429 | 0.025 | |||
-22.914*** | -0.402*** | -30.368*** | -0.233 | |||
-17.300*** | -0.118*** | 1.765 | 0.009 | |||
Adj. R2 | 0.186 | 0.045 | ||||
F stats | 68.93(0.000) | 4.60(0.001) |
注:考虑到相距较近的区域在同一时间的空气质量指数可能存在巨大差别,故自变量未纳入分析。 |
Tab. 2 Positively and negatively impacted regions on weekdays and weekends表2 工作日和非工作日受显著影响的社区数量分布 |
自变量 | 工作日 | 周末 | ||||||
---|---|---|---|---|---|---|---|---|
正相关 | 负相关 | 正相关 | 负相关 | |||||
数量 | 比例% | 数量 | 比例% | 数数量量 | 比例% | 数量 | 比例% | |
27 | 19 | 113 | 81 | 74 | 75 | 25 | 25 | |
35 | 38 | 56 | 62 | 18 | 53 | 16 | 47 | |
24 | 7 | 310 | 93 | 14 | 10 | 129 | 90 | |
37 | 63 | 22 | 37 | 65 | 64 | 36 | 36 | |
166 | 94 | 10 | 6 | 121 | 91 | 12 | 9 | |
None | 671 | 961 |
Fig. 1 Distributions of travel distance and duration on sunny and rainy days图1 晴雨天载客距离及行程时间分布 |
Tab. 3 Significantly impacted regions in urban area and suburban area表3 受自变量显著影响的主城区和郊区社区数量分布 |
自变量 | 工作日 | 周末 | ||||||
---|---|---|---|---|---|---|---|---|
主城区 | 郊区 | 主城区 | 郊区 | |||||
数量 | 比例/% | 数量 | 比例/% | 数量 | 比例/% | 数量 | 比例/% | |
118 | 14 | 22 | 4 | 81 | 10 | 18 | 3 | |
62 | 7 | 29 | 5 | 24 | 3 | 10 | 2 | |
279 | 33 | 55 | 9 | 116 | 14 | 27 | 4 | |
46 | 5 | 13 | 2 | 85 | 10 | 16 | 3 | |
152 | 18 | 24 | 4 | 108 | 13 | 25 | 4 | |
None | 193 | 23 | 478 | 76 | 436 | 50 | 525 | 84 |
合计 | 850 | 100 | 621 | 100 | 850 | 100 | 621 | 100 |
Fig. 2 Variations of travel flow between urban area and suburban area图2 工作日和周末晴雨时段主城区和郊区间出行流量变化分布 |
Fig. 3 Distributions of travel distance between urban area and suburban area on weekdays and weekends图3 工作日和周末晴雨时段主城区和郊区间载客距离分布 |
Fig. 4 Proportions of regions with different land use types impacted by raining图4 工作日和周末受下雨影响的用地类型分布 |
Fig. 5 Variations of travel flow between regions with different land use types on weekdays and weekends图5 工作日和周末晴雨时段不同用地类型间出行流量变化分布 |
Fig. 6 Distributions of travel distance from industrial, educational and residential area to commercial area on weekdays and weekends图6 工作日和周末工业、教育科研、居住用地至商业金融用地间载客距离分布 |
Fig. 7 Impacted regions by raining on weekdays and weekends图7 工作日和周末受下雨影响的社区空间分布 |
Fig. 8 Travel paths between International Garden and Optical Valley on sunny days and rainy days图8 晴天和雨天往返当代国际花园和光谷的载客轨迹分布 |
Fig. 9 Traffic volume, speed and duration between International Garden and Optical Valley on sunny days and rainy days图9 晴雨天当代国际花园至光谷间不同载客路线车辆比例、车速及时间分布 |
The authors have declared that no competing interests exist.
[1] |
|
[2] |
[
|
[3] |
[
|
[4] |
|
[5] |
|
[6] |
|
[7] |
[
|
[8] |
|
[9] |
[
|
[10] |
|
[11] |
|
[12] |
[
|
[13] |
|
[14] |
|
[15] |
|
[16] |
|
[17] |
Federal climate complex data documentation for integrated surface data. Available from
|
[18] |
[
|
[19] |
[
|
[20] |
|
[21] |
Analyzing 1.1 billion NYC taxi and Uber trips with a vengeance. Available from nalyzing 1.1 billion NYC taxi and Uber trips with a vengeance. Available from .
|
[22] |
[
|
[23] |
[
|
[24] |
|
[25] |
[
|
[26] |
Wuhan Land Resources and Planning Bureau. The comprehensive planning of Wuhan[J/OL]. .
|
[27] |
[
|
[28] |
WHO. Review of evidence on health aspects of air pollution - REVIHAAP Project 2013 (cited 21 May 2015)[EB/OL]. .
|
[29] |
|
[30] |
|
/
〈 | 〉 |