Redundant encoding is the use of more than one graphical or visual structure (e.g. color + position) to encode/represent one variable of data. For instance, in a 2D scatter plot, the x-axis values can be represented by both color+shape (e.g. red rectangles for series 1, blue circles for series 2 etc.). Actually, we can use a non-redundant code to represent these data (e.g. rectangles for series 1, circles for series 2). It is controversial whether redundant codes are effective or not. Some visualization theorists argue that certain redundant codes can increase accuracy and enhance perception.