城市海量手机用户停留时空分异分析——以深圳市为例
作者简介:徐金垒(1989-),男,硕士生,研究方向为时空数据的模式分析与推理。E-mail:xjlms@whu.edu.cn
收稿日期: 2014-05-20
要求修回日期: 2014-07-26
网络出版日期: 2015-02-10
基金资助
国家自然科学基金项目(41231171、41371420)
中国科学院资源与环境信息系统国家重点实验室开放基金(2013)
深圳市科技创新委基础研究“从大规模手机定位数据中挖掘居民出行的关键技术研究”(JCYJ20140610151856728)
The Spatio-temporal Heterogeneity Analysis of Massive Urban Mobile Phone Users’ Stay Behavior: A Case Study of Shenzhen City
Received date: 2014-05-20
Request revised date: 2014-07-26
Online published: 2015-02-10
Copyright
识别海量手机数据中蕴含的行为模式,是地理学的一个研究热点与难点。目前,较多研究针对手机用户移动特征开展,而对停留及其模式的研究则相对较少;其时空分异规律对理解城市人群动态,甚至优化城市系统至关重要。本文根据人们日常时空约束条件定义了手机用户停留,提出了基于海量手机位置数据的手机用户停留模式的提取方法,以深圳市约790万个匿名手机用户一天的海量手机位置数据为例,识别出了覆盖约98%用户的典型停留模式,并结合该城市土地利用的空间分布与分异特征,剖析不同停留模式的手机用户空间分异特征和城市不同区域停留次数的时段分异特征。研究发现:(1)15种停留模式可覆盖约98%的手机用户,而且其一天不同的停留位置数量不超过4个;(2)15种停留模式手机用户在城市区域空间上的分布存在分异现象,严重受制于土地利用的空间分布;(3)城市不同区域停留次数的时段分异特征与该区域常住人口、人口密度,以及区域主要职能和性质存在较强的相关性。研究结论对理解城市手机用户行为模式的群体特征有积极的意义,对城市土地利用的科学决策和城市交通规划与预测有重要参考价值。
徐金垒 , 方志祥 , 萧世伦 , 尹淩 . 城市海量手机用户停留时空分异分析——以深圳市为例[J]. 地球信息科学学报, 2015 , 17(2) : 197 -205 . DOI: 10.3724/SP.J.1047.2015.00197
Identifying human behavior patterns embedded in massive mobile phone data records is a research hotspot and difficulty of study in geography. While lots of researches are aiming at investigation of the mobility characteristics of mobile phone users, studies focusing on the stay of mobile phone users and their stay patterns are relatively little. The spatio-temporal heterogeneous regularity is vital for understanding the urban human dynamic and optimizing urban system. Therefore, this paper explored the stay patterns of mobile phone users in order to understand the human mobility patterns and their features. According to the spatial and temporal constraints of human routine and travel behavior, this paper defined the stay conditions and the stay patterns of urban mobile phone users, and come up with methods of identifying and extracting stay patterns from massive mobile phone data records. Then, we took about 7.9 million mobile phone users’ one day data records in Shenzhen City in China as an example, processed them with the designed method, and identified some typical stay patterns which covered about 98% of the urban mobile phone users. In addition, we conducted statistical analysis for urban mobile phone users’ stay patterns and analyzed their spatial distribution with respect to different administrative districts combining with the city’s land use pattern. Furthermore, we dissected the heterogeneous characteristics of the spatial distribution of different stay pattern’s users and the stay frequencies against different administrative districts over a 22-hour period. The study shows: (1) 15 types of stay patterns can cover 98% urban mobile phone users’ behavior and their stay locations are no more than 4 in one day. The stay patterns with 2 stay locations shows the highest probability that nearly half of the mobile phone users have this tendency. (2) The spatial distribution of the 15 stay patterns in the city is evidently heterogeneous, which is subjected to the urban land use. We found that the higher proportion of leisure and entertainment land in a certain administrative district, the stay patterns with more stay locations in this region are likely to occur. (3) The heterogeneity of the stay frequency for different urban administrative district is strongly related to its permanent resident population, population density and major function of the district. The study conclusions are useful in understanding the population characteristics of urban phone users’ behavior, the activity patterns and urban residents’ commuting behavior. Moreover, it can provide important references for scientific decisions of urban land use, urban transportation planning and prediction.
Fig. 1 The stay points of one mobile phone user图1 手机用户停留点 |
Tab. 1 The instance of one phone user's records data表1 手机用户数据记录实例 |
时间 | 经度(long) | 纬度(lat) |
---|---|---|
0:50:14 | 113.×××× | 22.×××× |
1:50:11 | 113.×××× | 22.×××× |
...... | ||
22:10:01 | 113.×××× | 22.×××× |
23:09:56 | 113.×××× | 22.×××× |
Tab. 2 The statistical results of 15 stay patterns表2 15种停留模式统计结果 |
停留模式 | 手机用户数 | 所占比例(%) |
---|---|---|
SP2_3 | 1 502 666 | 19.06 |
SP1_1 | 1 162 855 | 14.75 |
SP2_4 | 995 836 | 12.63 |
SP1_2 | 808 152 | 10.25 |
SP2_2 | 622 709 | 7.90 |
SP3_4 | 559 477 | 7.10 |
SP1_3 | 444 842 | 5.64 |
SP2_5 | 383 145 | 4.86 |
SP3_3 | 379 625 | 4.81 |
SP3_5 | 332 651 | 4.22 |
SP1_4 | 142 039 | 1.80 |
SP4_4 | 105 119 | 1.33 |
SP4_5 | 104 532 | 1.33 |
SP3_6 | 101 876 | 1.29 |
SP2_6 | 87 971 | 1.12 |
总计 | 7 733 495 | 98.09 |
Tab. 3 The statistics of stay locations’ classification表3 停留位置分类统计表 |
停留位置数 | 手机用户数(人) | 比例(%) |
---|---|---|
1 | 2 557 888 | 32.44 |
2 | 3 592 327 | 45.57 |
3 | 1 373 629 | 17.42 |
4 | 209 651 | 2.66 |
总计 | 7 733 495 | 98.09 |
Fig. 2 The administrative district and land use map of Shenzhen City图2 深圳市行政区划和土地利用分布图 |
Fig. 3 The spatial distribution for phone users of the 15 stay patterns图3 15种停留模式手机用户空间分布 |
Tab. 4 The proportion statistics of the 15 daily stay patterns in different administrative districts表4 15种停留模式在不同行政区中的比例 |
停留模式 | 不同行政区的比例(%) | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
宝安区 | 南山区 | 福田区 | 罗湖区 | 盐田区 | 龙岗区 | 龙华新区 | 光明新区 | 坪山新区 | 大鹏新区 | |
SP2_3 | 24.85 | 12.70 | 18.46 | 8.66 | 1.35 | 14.98 | 3.45 | 12.12 | 2.67 | 0.76 |
SP1_1 | 25.76 | 9.82 | 13.58 | 7.88 | 1.24 | 19.64 | 4.04 | 13.95 | 3.00 | 1.10 |
SP2_4 | 28.97 | 11.30 | 13.29 | 6.21 | 1.27 | 17.37 | 4.27 | 12.85 | 3.51 | 0.95 |
SP1_2 | 27.15 | 10.31 | 13.27 | 7.47 | 1.21 | 18.48 | 4.23 | 13.90 | 3.09 | 0.90 |
SP2_2 | 23.40 | 12.22 | 18.39 | 11.48 | 1.22 | 15.55 | 3.18 | 11.67 | 2.22 | 0.66 |
SP3_4 | 24.12 | 13.68 | 18.83 | 9.02 | 1.25 | 14.56 | 3.17 | 11.65 | 2.85 | 0.86 |
SP1_3 | 28.92 | 10.36 | 12.20 | 6.46 | 1.13 | 18.47 | 4.66 | 13.59 | 3.33 | 0.89 |
SP2_5 | 30.78 | 10.16 | 10.24 | 4.77 | 1.25 | 19.81 | 4.68 | 13.03 | 4.16 | 1.11 |
SP3_3 | 21.37 | 14.12 | 21.56 | 11.65 | 1.16 | 13.63 | 2.71 | 11.01 | 2.12 | 0.68 |
SP3_5 | 27.33 | 12.50 | 14.73 | 6.48 | 1.28 | 16.78 | 3.80 | 12.30 | 3.75 | 1.05 |
SP1_4 | 30.33 | 9.96 | 11.27 | 5.78 | 1.04 | 18.90 | 5.04 | 13.42 | 3.44 | 0.81 |
SP4_4 | 19.74 | 15.43 | 23.40 | 12.02 | 1.14 | 12.83 | 2.40 | 10.25 | 2.11 | 0.69 |
SP4_5 | 22.43 | 15.21 | 20.12 | 9.37 | 1.17 | 13.96 | 2.79 | 11.16 | 2.96 | 0.85 |
SP3_6 | 29.46 | 11.17 | 11.35 | 5.13 | 1.30 | 18.76 | 4.32 | 12.85 | 4.54 | 1.12 |
SP2_6 | 30.95 | 9.62 | 8.49 | 4.24 | 1.24 | 21.91 | 4.49 | 13.19 | 4.75 | 1.12 |
Tab. 5 The proportion statistics of mobile phone users with stay behavior in different administrative districts表5 不同行政区停留手机用户数比例 |
行政区 | 宝安区 | 南山区 | 福田区 | 罗湖区 | 盐田区 | 龙岗区 | 龙华新区 | 光明新区 | 坪山新区 | 大鹏新区 |
---|---|---|---|---|---|---|---|---|---|---|
停留手机用户数比例(%) | 25.74 | 12.17 | 16.35 | 8.26 | 1.25 | 16.35 | 12.35 | 3.66 | 3.01 | 0.87 |
Fig. 4 The statistics of land use distribution in different administrative districts图4 不同行政区土地利用分布统计图 |
Fig. 5 The stay frequency in different administrative districts over 22 hours图5 不同行政区的停留次数随时段变化 |
Fig. 6 The statistics property of different administrative districts图6 不同行政区统计属性 |
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] |
2013年深圳统计年鉴.
|
[17] |
|
/
〈 | 〉 |