Understanding the Log Scale

In this lecture, we are going to explore the log scale. It is another scale that solves a specific problem that other scales may not be able to address. Let’s explore the problems this scale can solve.

The problem

The log scale transforms continuous data into another set of continuous data. This is the same as the linear scale, so what makes it different?

The log scale will apply a logarithmic transformation on the domain before transforming the data. The purpose of a log transformation is to make the distribution of data appear normal. If you are not familiar with what a log transformation is, that is perfectly fine. The scale does most of the work. Let’s briefly discuss why and when you would want to use a log scale.

Throughout your career, you will be dealing with different types of data. If you are dealing with data that relates to money, time, or distance, a log scale may be necessary. This rule is not absolute. It is more of a recommendation. Generally speaking, these are the most common types of data that a log scale is used for.

Why log scales

That answers what kind of data a log scale can be used for, but not why we should use a log scale. Let’s pretend we had a dataset that had the numbers 1, 1 million, and 1 billion.

What do you notice about these numbers? Their distribution varies greatly. The distance between 1 and 1 million is small compared to the distance between 1 million and 1 billion.

Let’s say we had to draw this data.