Chernoff face

Chernoff faces are developed by Herman Chernoff for visualization of multivariate data by human faces method.

The basic idea of data visualization by faces is that it falls a people easy to recognize even small differences in human faces, and thereby the complexity of multivariate data (eg properties of beer, variables of apartment dwellings, etc.) through this system of Chernoff - "Faces " easier to grasp.

Each variable of a data set is thereby assigned to one of eighteen facial features, such as position, position or size of the face, eyes, eyebrows, mouth and nose, creating a unique look for each observation.

Construction of a Chernoff face

In this picture is shown an example of how a construction of a Chernoff face is performed. All lengths marked with letters are determined by the values ​​of an observation. Here, these values ​​must be subjected to a previous normalization.

Asymmetry

1981 beat Bernhard Flury and Hans Riedwyl before asymmetric Chernoff faces. As a Chernoff face is vertically symmetrical, so the left and right half of the face are identical, can be omitted on one side of the face in principle. This redundancy in the representation also criticized Tufte. By the use of asymmetric faces, the number of variables can be displayed can be doubled.