CI Thermometer: Visualizing Confidence Intervals in Correlation Analysis

IEEE Comput Graph Appl. 2017 Nov-Dec;37(6):103-108. doi: 10.1109/MCG.2017.4031058.

Abstract

Correlation analysis is one of the most popular statistical methods. Despite that, the way research reports correlation is often misleading. The difficulty increases with the amount of information that must be read and interpreted. The proposed CI thermometer makes correlation matrices much easier to read and provides information that would be difficult to interpret when presented in another way.