Probability Distributions › Introduction

Every data scientist encounters probability distributions from the earliest stages of exploratory data analysis through the most sophisticated machine learning pipelines. Understanding distributions enables practitioners to make informed assumptions about data-generating processes, select appropriate statistical tests, and interpret model outputs with confidence.

When we observe data in the real world (whether it's customer purchase amounts, manufacturing defect rates, or time between system failures), we often find that these observations follow recognizable patterns. Some data clusters symmetrically around a central value like a bell curve, while other data exhibits pronounced skewness or discrete jumps. Probability distributions provide the mathematical language to describe these patterns precisely, connecting theoretical models to empirical observations.

This chapter explores the fundamental probability distributions that every data scientist should know. We begin with the normal distribution, the most ubiquitous distribution in statistics, before examining distributions suited to specific data types and contexts. By understanding when and why to apply each distribution, practitioners can build more accurate models and draw more reliable conclusions from data.