Data Visualization in R and Python

11Heatmaps

Heatmaps are a type of graphic that is usually easy to produce and could be aesthetically pleasant and effective to convey information in an intuitive way. In practice, what a heatmap shows is a color-based representation of a data frame in rectangular form, with two categorical variables associated to the sides of the heatmap (corresponding to the Cartesian axes), and a third variable, either continuous or categorical, whose values are converted into a color scale. The idea is that, through the color representation, an observer could easily and intuitively grasp the values of the third variable corresponding to the two variables on the axes. The information conveyed by a heatmap is largely qualitative, the color scale usually has quantitative values but, especially with a continuous gradient, the exact value associated to a certain hue is difficult to determine, so what an observer gets is often a broad approximation of the real value. Therefore, with respect to the corresponding data frame, a heatmap is certainly less precise but it gains in simplicity for an observer to get the informational content. In addition to this, heatmaps, being colorful and with their regular structure, are well-adapted to be used in creative ways and combined with different graphical elements.

Dataset

In this chapter, we make use again of data from Bicycle thefts in Berlin (transl. Fahrraddiebstahl in Berlin) from the Municipality of Berlin, Germany, Berlin Open Data, previously introduced. ...