基于双标图、三元图与平衡树状图的中国居民消费结构变化的多元可视化分析

Multivariate Visualization Analysis of Changes in the Consumption Structure of Chinese Residents Based on Double-Standard Graphs, Ternary Graphs and Balanced Tree Graphs

摘要: 居民消费结构是衡量经济发展质量的重要指标，但消费数据作为成分数据具有“定和约束”。本文对2014~2023年中国31个省级行政区八类居民消费支出成分数据，分别运用双标图、三元图和平衡树状图三种可视化方法，系统分析了消费结构的时空演变特征。双标图揭示了消费品类间的关联模式与省份聚类；三元图直观展示了2023年各省消费结构的整体分布，并通过两种分类方案揭示出时间与空间的聚类模式；平衡树状图量化了高升级型与刚需型两组省份在各类消费上的均值与方差差异。三种方法相互印证，为深入理解中国居民消费变迁提供了多维视角。

Abstract: The consumption structure of residents is an important indicator for measuring the quality of economic development. However, consumption data, as compositional data, is subject to the “constant sum constraint”. This paper systematically analyzes the spatio-temporal evolution characteristics of the consumption structure of eight types of resident consumption expenditures in 31 provinces of China from 2014 to 2023 by using three visualization methods: biplot, ternary plot, and balanced dendrogram. The biplot reveals the association patterns among consumption categories and the clustering of provinces; the ternary plot visually presents the overall distribution of consumption structures in all provinces in 2023 and reveals the clustering patterns in time and space through two classification schemes; the balanced dendrogram quantifies the mean and variance differences in various consumption types between the high-upgrading and essential-demand groups of provinces. These three methods complement each other, providing a multi-dimensional perspective for a deeper understanding of the changes in Chinese residents’ consumption.