Practical Predictive Analytics

Interval data

This is numeric data in which you can measure distances but cannot take ratios. Temperature data and calendar data are two examples of interval data. All of the arithmetic differences between two consecutive data points on the numeric scale are measured exactly the same way, so you can take meaningful differences.

Often, interval data is treated as ratio data. This is often seen when performing calculations such as averages on likert scales (even though this is not technically correct). With ordinal data, you can compute frequencies, calculate the median, and compute percentiles but you cannot do any arithmetic or calculate a mean.