测绘通报 ›› 2022, Vol. 0 ›› Issue (6): 93-97.doi: 10.13474/j.cnki.11-2246.2022.0178.

• 学术研究 • 上一篇    下一篇

利用腾讯位置大数据进行多尺度人口空间化估算

李慧敏, 罗大伟, 窦世卿   

  1. 桂林理工大学测绘地理信息学院, 广西 桂林 541000
  • 收稿日期:2021-08-09 发布日期:2022-06-30
  • 通讯作者: 窦世卿。E-mail:doushiqing@glut.edu.cn
  • 作者简介:李慧敏(2000-),女,研究方向为GIS技术及应用。E-mail:lihuimin_glut@qq.com
  • 基金资助:
    国家自然科学基金(42061059);广西自然科学基金(2020JJB150025)

The estimation of population on multi-spatial scale using Tencent location big data

LI Huimin, LUO Dawei, DOU Shiqing   

  1. College of Geomatics and Geoinformation, Guilin University of Technology, Guilin 541000, China
  • Received:2021-08-09 Published:2022-06-30

摘要: 社会感知数据的多样性,为人口估算方法的研究引入了新数据和视角。本文利用腾讯位置大数据进行多尺度人口空间化估算研究,以建设用地和住宅小区体积数据作为相关因子,将5 km格网精细化到1 km及住宅小区级尺度,通过广西壮族自治区、桂林市、临桂区3个不同尺度分析腾讯位置大数据与人口统计数据的相关关系,利用多项式回归模型分析得到3种尺度的人口空间化模型,并对模型的正确性进行验证。结果表明,省级尺度的人口估算模型验证精度R2达0.984 5,市级尺度的模型验证精度R2达0.9770;小区精细尺度的模型验证精度R2达0.979 9,估算结果可信度高。

关键词: 腾讯位置大数据, 人口估算, 多尺度, 格网化

Abstract: The diversity of social perception data brings new data and perspectives to the study of population estimation methods. In this paper, the multi-scale population spatial estimation is studied by Tencent location big data. With the volume data of construction land and residential area as the relevant factors, the 5 km grid is refined to 1 km and residential area level scale. By analyzing the correlation between Tencent location big data and demographic data at three different scales of Guangxi, Guilin and Lingui district, the spatial model of population at three scales is obtained by polynomial regression model analysis, and the correctness of the model is verified. The accuracy of the population estimation model is 0.984 5 on the scale of provincial level and 0.977 0 on the scale of city level, the verification accuracy is 0.979 9 in terms of community scale. The estimation results have high reliability.

Key words: Tencent location big data, population estimation, multi-scale, grid transformation

中图分类号: