Practical Predictive Analytics

Ratio data

The highest level of numeric data is ratio scale. With ratio data, you can make meaningful comparisons using division and subtraction. It is meaningful to say that you have acquired twice as many customers when you went from 50 to 100 customers in one month. Ratio data also adds 0 to the domain of numbers (which interval data does not contain), so it is meaningful to say that you have 0 customers (although not desireable). Weight and income are other examples of ratio data. however, the fahrenheit and celsius temperature scale do not have a 0 meaning no temperature. That creates some interesting comparisons if you would try to make temperature comparisons. For example, in fahrenheit, a temperature of 36/18 equals 2, but the equivalent ...