Circular data models

Steve Simon

1999-12-15

Categories: Blog post Tags: Circular data

When your data has a cyclic or circular component, then the use of basic trigonometry can provide a statistical model that allows for cyclic effect. Examples of cyclic components are time of the day and wind direction.

The unique feature is that data at opposite ends of the interval are actually close to one another. For example

An average for cyclic data is pretty much useless whenever you have data at the extremes. For example

You can find an earlier version of this page on my original website.