Why Use Ternary Graph?
STATISTICA has a rich suite of data visualizations. In recent weeks, StatSoft has received questions on the Ternary Graph (also known as Trilinear). So we thought a high-level overview would be helpful.
This graph is highly useful in science and manufacturing. Maybe you need to:
The ternary graph is made to explore relationships between three or more dimensions. And this relationship is constrained. The sum of these parts must be a constant like 42 or 100%.
You graph component X, component Y, component Z and response V1, V2, etc…
The math behind this graph type requires that the “sum of parts” is equal to 1. Because data isn’t perfect…. the real world rarely adds up to 1… STATISTICA will automatically rescale proportions so the parts will sum to 1. This means less work for you.
If you don’t want this automatic rescaling, you have a couple of other options. You can recode to “pseudo components” (Cornell, 1990). You can ignore cases that have invalid sums. You can filter the “bad” data out.
StatSoft has 6 different types of Ternary like Scatterplot, Surface, Contour and Trace. The options can be viewed on the Advanced tab.