基于大规模手机定位数据的群体活动时空特征分析
作者简介:曹劲舟(1991-),男,湖南益阳人,博士生,主要从事时空大数据分析与挖掘研究。E-mail:caojinzhou@whu.edu.cn
收稿日期: 2016-11-04
要求修回日期: 2016-12-24
网络出版日期: 2017-04-20
基金资助
国家自然科学基金项目(41401444、41371377、41671387)
深圳大学青年教师科研启动项目(2016065)
国土资源部城市土地资源监测与仿真重点实验室开放基金资助课题(KF-2016-02-009)
Spatio-temporal Analysis of Aggregated Human Activities Based on Massive Mobile Phone Tracking Data
Received date: 2016-11-04
Request revised date: 2016-12-24
Online published: 2017-04-20
Copyright
城市空间与居民行为不断交互,相互影响。探究城市空间中的群体活动分布及其时空变化能够帮助数据驱动的城市规划与城市治理。基于大数据的时空间群体活动研究是当前时空大数据研究的一个热点。本文以深圳市为例,基于约1000万手机用户在某一工作日的基站尺度的手机定位数据,识别用户停留位置和停留活动,重建活动语义信息,分析用户的停留点和停留活动的分布差异,研究群体活动的时空分布模式,探讨人群活动模式的多样分布特征。研究表明:停留位置和活动分布存在差异,每人每天平均的停留个数约为2.1个,而每人每天平均从事的活动约为3.4个;不同类型的活动在时间上存在波动;群体活动存在空间分异特征,整体上服从“空间幂律”。本研究揭示了城市空间中群体活动的多样性及其时空分布特征,对于城市居民活动研究、城市交通优化和城市规划具有重要的意义。
曹劲舟 , 涂伟 , 李清泉 , 曹瑞 . 基于大规模手机定位数据的群体活动时空特征分析[J]. 地球信息科学学报, 2017 , 19(4) : 467 -474 . DOI: 10.3724/SP.J.1047.2017.00467
Urban space and the behavior of human activities constantly interact with each other. Investigation on distribution of aggregated human activities and spatio-temporal change benefits data-driven policy-making in urban planning and urban governing. In the era of big data, with the development of information and communication technologies, it is possible to collect city-scale data with high resolution in space and time by various location-aware devices and sensors. Exploration of spatial-temporal activities attracts a lot of attention. By taking about 10 million one-day tracking data of mobile phone users in Shenzhen, China as an example, this paper firstly identified their stay locations according to spatial and temporal rules to generate stay trajectory for each individual and recovered activity semantic information by labelling activity types for each stay locations. Then, the significant differences in patterns of distributions of stay locations and their activities were analyzed. Spatial and temporal distributions of different human activities were explored, respectively. The study shows that the distribution of stay locations and activities is obviously heterogeneous. The average number of stay locations of an individual per day is 2.1, while the average number of activities an individual engaged in per day is 3.4. This study furthermore suggests that different types of activities have temporal variance and spatial heterogeneity. The temporal distribution fluctuates significantly over 24 hours, which is in accordance with daily routine. The spatial distribution overall obeys “space power law”, and the spatial distribution of social activity, which has a faster-down tail, shows a more obvious pattern of spatial segregation than the other two activities. The study revealed the diversity and heterogeneity of spatial and temporal distribution of human aggregated activities in urban space, which is meaningful in analyzing human activities research and facilitating urban traffic optimization and urban planning.
Fig. 1 The study area图1 研究区域 |
Tab. 1 Examples of mobile phone location records表1 移动手机定位数据示例 |
用户id | 小时数 | 基站id | 基站经度 | 基站纬度 |
---|---|---|---|---|
536**** | 0 | 19** | 114.14** | 22.60** |
536**** | 1 | 19** | 114.14** | 22.60** |
536**** | 2 | 54** | 114.12** | 22.58** |
… | … | … | … | … |
536**** | 23 | 14** | 114.14** | 22.60** |
注:为了隐私保护,具体数值以*号标识 |
Fig. 2 Framework of spatio-temporal analysis of aggregated human activities using cellphone location data图2 基于大规模手机定位数据的群体活动时空特征分析流程图 |
Fig. 3 Cumulative distribution of the distance between any adjacent base tower stations图3 任意相邻基站距离累计分布 |
Fig. 4 Identification of home, work and social activities from stay points trajectory图4 从停留点轨迹中识别家庭-工作-社会活动方法 |
Fig. 5 The statistics of daily stay points and activities图5 一天内停留位置和停留活动统计分布 |
Fig. 6 Daily temporal distribution of number of stay points图6 停留个数不同时间段分布 |
Fig. 7 Percentage of population by five types of detection of home and work locations图7 5种类型的居家和工作位置识别结果的人口比例 |
Fig. 8 Correlation between the spatial distributions of population based on home locations at street level and the population distribution from 2010 census data图8 街道级别居家位置人口分布与2010人口普查分布比例相关性分析 |
Fig. 9 Daily temporal distribution of the volume of total activity, work activity, social activity and home activity图9 总活动量和工作、社会、家庭活动量不同时间段变化分布 |
Fig. 10 Spatial distribution of activity density图10 总体活动密度空间分布 |
Fig. 11 Spatial distributions of densities of home, work and social activities图11 家庭、工作、社会活动密度空间分布 |
Fig. 12 Complementary CDF of ranks of activity densities normalized by the mean in different categories图12 不同类别活动密度排名的互补累积分布 |
Fig. 13 The statistics of distribution of activity volumes in different administrative districts图13 不同行政区活动量分布统计 |
Fig. 14 The statistics of distribution of activity densities in different administrative districts图14 不同行政区活动密度分布统计 |
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] |
|
[18] |
|
[19] |
|
[20] |
深圳市人民政府公报2016年第39期(总第983期)[R].
|
[21] |
深圳市2015年全国1%人口抽样调查主要数据公报[EB/OL]..
|
[22] |
[
|
[23] |
|
/
〈 | 〉 |