Bicomponent Trend Maps: A Multivariate Approach to Visualizing Geographic Time Series

Area of Science:

• Geographic Information Science
• Spatial Analysis
• Data Visualization

Background:

• Traditional temporal mapping methods struggle to represent complex spatio-temporal patterns across diverse regions and timeframes.
• Effective visualization of long-term trend variations is crucial for understanding demographic shifts.

Purpose of the Study:

• Introduce bicomponent trend mapping as an alternative approach for illustrating spatio-temporal patterns.
• Demonstrate the utility of bicomponent trend mapping for analyzing population trends in U.S. urban cores from 1950 to 2000.

Main Methods:

• Employ principal component analysis (PCA) to identify key dimensions of trend variations.
• Utilize bivariate choropleth mapping to visualize two distinct dimensions of long-term trends.
• Develop a bicomponent trend matrix for interpreting trend types and visualizing principal components.

Main Results:

• Bicomponent trend mapping effectively illustrates two dimensions of long-term trend variations.
• The bicomponent trend matrix serves as a legend and visualization tool for principal components.
• Application to U.S. urban population trends reveals interpretable relationships among trend classes.

Conclusions:

• Bicomponent trend mapping offers a novel method for visualizing spatio-temporal data, particularly for demographic trends.
• While not depicting as wide a variety of properties as other multivariate methods in static displays, it enhances interpretability of trend relationships.
• The approach provides unique classification flexibility, beneficial for interactive data exploration environments.