Transparency is a useful dimension for attracting the viewer’s attention toward certain points and away from other points.
Transparency allows all feature types and is good with numerical features given the continuous nature of the representation. Transparency is also good with binary categorical features (categorical with two categories) — this dimension is a great way to highlight a small category in a crowded visualization, such as outliers. While transparency can also be used with larger categorical features, it is not recommended since it is difficult to differentiate between multiple levels of transparency.
Transparency settings are found by switching to the Others tab on the Mapping panel :
- Normalization – Numerical only, see Normalization
- Scale – Adjusts the size of all points to be either bigger or smaller. Note: This setting can be changed even when there is no feature mapped to transparency. By default, all points are slightly transparent — this is to allow the viewer to see when multiple points are overlapping each other. Points can be made entirely opaque by turning up this setting, which can be useful to make the visualization more vibrant for presentations.
- Range – Numerical only, filters out points outside of the two selected values