Data Science Programming All-in-One For Dummies

Chapter 2

Considering Outrageous Outcomes

IN THIS CHAPTER

Defining when an outcome is outrageous

Detecting outliers

Using the simple univariate method

Using the multivariate approach

If you work with data long enough, you eventually start to gain an appreciation for when the output of an analysis looks right. It may not be the output you expected, but when you start thinking about it, the output is consistent with the data — it makes sense. Unfortunately, the output you receive might not always make sense, and that’s when the output becomes outrageous. You start seeing results like the sun coming out at midnight and the anticipated income from a new store being well into the negative numbers. Of course, recognizing outrageous isn’t always so easy, so the first part of this chapter begins with defining outrageous.

An outlier is data that lies outside the expected range. It’s an indicator that something may be wrong with your data or the method used to analyze it. Outliers can skew the results of an analysis or they can indicate that your original assumptions are incorrect. In some ...