Heat Maps Defined
June 24, 2008
Heat maps are an excellent way to visualise data the reader instinctively knows what the information means.
However in trying to answer the questions “what is a heatmap?”, I found definitive information hard to find.
I eventually came across an article by Stephen Few which not suprisingly gives a clear and concise definition and illustrates this with some real life examples.
Multivariate Analysis Using Heatmaps
By Stephen Few
Published: October 10, 2006
This is the third article in a series that began in July with the article entitled, “An Introduction to Visual Multivariate Analysis.”
Prior articles in this series have examined how table lens and parallel coordinates displays can be used to explore and analyze multivariate information. In this article, I describe the use of multivariate heatmap matrices.
In general, the term heatmap refers to any display that uses color to represent quantitative data. We are all familiar with heatmaps in the form of weather maps, which use color to encode values such as temperature or rainfall. Heatmaps also come in forms other than geographical maps. When heatmaps are used to encode multivariate data – several variables that measure different aspects of some set of entities (for example, customers, countries, or products) – they are usually structured as a matrix of columns and rows. Figure 1 is a multivariate heatmap matrix, which displays a separate employee per row (the entities) and a separate measure per column (the variables). In this case, the heatmap’s purpose is to help us determine what factors most influence employee job satisfaction, which appears in the leftmost column labeled Working Conditions. Employees were asked to rate their working conditions as Very Poor (the lightest color), Poor, Acceptable, Good, or Very Good (the darkest color). Each of the other variables (Salary, etc.) has been encoded as a continuous range of grayscale colors, ranging from the lightest for the lowest value through the darkest for the highest value. By examining a single row, you can see a particular employee’s complete multivariate profile. By scanning a column, you can see the complete set of values for a particular variable across all employees, such as the average number of hours they work per week (the third column).
All trademarks are the properties of their respective owners
