Course contentsShow
← Data Science
Lesson 179 of 2,1455. Continuous Probability DistributionsFree lesson

When Variables Are Log-Normally Distributed

Real-world phenomena like stock prices, incomes, and biological measurements that follow log-normal patterns.

When Variables Are Log-Normally Distributed

You'll learn to recognize real-world scenarios where log-normal distributions naturally appear.

The Core Idea

While many phenomena follow a normal (bell-curve) distribution, some real-world variables cannot be negative and tend to have a long right tail—think extreme values are possible but rare. When you take the natural logarithm of these variables and they become normally distributed, the original variable is log-normally distributed.

Common Real-World Examples

Stock Prices and Investment Returns

Stock prices can never drop below zero, but they can theoretically grow without limit. Over time, they tend to follow log-normal patterns because percentage changes compound multiplicatively, not additively.

Income Distribution

Most people earn moderate incomes, but a small number earn extremely high salaries. This creates the characteristic right-skewed shape of the log-normal distribution. Income cannot be negative, and the highest earners stretch far to the right.

Biological Measurements

Many biological variables are log-normally distributed:

  • Cell sizes in organisms
  • Blood pressure readings across populations
  • Latency periods for disease incubation
  • Survival times in medical studies

File Sizes and Network Traffic

In technology, file sizes on servers and data packet sizes in network traffic often follow log-normal distributions—most files are small, but occasionally you encounter very large ones.

Why This Matters

Recognizing log-normal patterns helps you choose the right analytical tools. If you mistakenly treat log-normal data as normal, your probability calculations and predictions will be wildly incorrect, especially in the tails.

Key Takeaway: Log-normal distributions appear naturally when variables are always positive, result from multiplicative processes, and have long right tails—common in finance, biology, income, and technology data.