UCSC-SOE-08-24: Multivariate Visualization of Socio-Economic Indicators using Geographic Maps

Mike Patterson and Suresh K. Lodha
11/26/2008 09:00 AM
Computer Science
National data on several socio-economic indicators are being collected and disseminated by various agencies including World Bank and United Nations. These indicators are typically visualized using well-known bivariate techniques including geographic maps, line graphs, bar graphs, and scatter plots, with some notable exceptions such as the visualization of four variables on a scatter plot by adding glyph sizes and colors. Judicious use of multivariate visualization of several global indicators may allow useful insights.

In this work, we introduce a novel visualization technique that allows pseudo-coloring of the boundary of a country based on an indicator. Boundaries of neighboring countries are separated using a displacement method along normals. This visualization coupled with three standard techniques of pseudo-coloring a country, glyph size, and glyph color, and judicious use of opacity allows us to display four variables simultaneously on a geographic map effectively. Working with sociologists, we applied our techniques to investigate global socio-economic indicators available in reliable databases. We have built a flexible, interactive, real-time visualization system that allows extensive investigation of these indicators by sociologists and computer scientists working together. Flexibility of our visualization platform allows us to map a fifth variable using cartograms. Our system also supports real-time animation of indicators over time. We have demonstrated that the multivariate visualization of global indicators provides valuable insights in many cases not easily available otherwise.

UCSC-SOE-08-24