HOTPERIODS: Visual Correlation Analysis of Interval Data

With the ever increasing amount and complexity of data, visual analysis becomes a fundamental tool to spot correlations and other relationships in data. Most of the previous techniques (e.g., scatter plots or heatmaps) focus on point data, i.e., data with point measures, such as prices or volumes. In this demo paper, we focus on data with interval measures, that is data where measures consist of an interval or range of values, such as price ranges or time intervals. We present a tool, termed HOTPERIODS, which allows to visualize correlations between two interval measures in the two-dimensional space, where the two measures represent a rectangle. To visualize such data, we first perform a rectangle aggregation. The result of this aggregation is a density matrix, where each cell stores the number of rectangles that cover the corresponding points in space. For the visualization of the density matrix, color-coding is used to represent different density values similar to heatmaps. We illustrate the usefulness of HOTPERIODS for the analysis of stock market data and tourism data, both of which show interval measures.