地理学报 ›› 2016, Vol. 71 ›› Issue (4): 564-575.doi: 10.11821/dlxb201604003

• • 上一篇    下一篇

社会感知视角下的若干人文地理学基本问题再思考

刘瑜   

  1. 北京大学遥感与地理信息系统研究所,北京100871
  • 收稿日期:2015-08-24 发布日期:2020-05-22
  • 作者简介:刘瑜(1971-), 男, 山东诸城人, 教授, 博士生导师, 中国地理学会会员(S110007302M), 主要研究方向为地理信息科学。E-mail: liuyu@urban.pku.edu.cn
  • 基金资助:
    国家自然科学基金项目(41271386, 41428102); 资源与环境信息系统国家重点实验室开放基金

Revisiting several basic geographical concepts:A social sensing perspective

LIU Yu   

  1. Institute of Remote Sensing and Geographical Information Systems, Peking University, Beijing 100871, China
  • Received:2015-08-24 Online:2020-05-22
  • Supported by:
    National Natural Science Foundation of China, No.41271386, No.41428102; State Key Laboratory of Resources and Environmental Information System]

摘要: 近年来,不同类型大数据在地理研究中得到了越来越多的重视,许多学者基于手机、社交媒体、出租车等数据开展了大量实证研究。社会感知概念刻画了地理空间大数据基于大量人的行为时空模式获取地理环境特征的的技术手段,该手段有助于重新审视地理学研究中的一些基本问题,因而本文选择了空间分布和空间交互这两个基本地理概念以及定性方法和定量方法这两个人文地理基本研究方法展开讨论。大数据从微观个体和宏观群体两个层面同时感知空间分布和空间交互,可以定量分析其中的距离以及尺度效应。进而,由于小样本访谈人群和场所是定性研究的基础,而大数据可以通过定量方法识别特定人群和场所并进行刻画,因此,社会感知手段为集成定性和定量研究方法,构建混合地理学奠定了基础。

关键词: 大数据, 社会感知, 空间分布, 空间交互, 定性方法, 定量方法

Abstract: Recently, various big data are drawing more and more attention in geographical research and many scholars have conducted lots of empirical studies using mobile phone data,social media data, taxi data, and so forth. Social sensing,a newly proposed concept, represents the capability of revealing socio- economic geographical features by capturing the spatial behavior patterns of a large population. Given that the term "environment" in humanenvironment interaction studies have involved the behavioral environment, social sensing techniques provide us a new approach to understanding human- environment interactions.Additionally, the emergence of social sensing helps us to rethink several fundamental issues in geographical studies. This article revisits two groups of core concepts: spatial distribution and spatial interaction, as well as qualitative method and quantitative method. Based on the fact that big data measure distributions and interactions at both individual and aggregate levels, we can quantify the underlying distance and scale effects from the observed patterns. To tackle space and population heterogeneity, clustering methods can be introduced to decompose a space and/ or a population into relatively homogeneous human groups and places. Considering that human groups and places are essential to qualitative studies, we argue that social sensing offers an opportunity to integrate big data and survey-based small data, and consequently, qualitative and quantitative methods are integated. Obviously, the second merit makes it possible to construct hybrid geography.

Key words: big data, social sensing, spatial distribution, spatial interaction, qualitative method, quantitative method