Compute arithmetic mean, geometric mean, harmonic mean, median, mode, variance, standard deviation, and more from any data set. Visualize your data distribution with an interactive chart.
In statistics, the mean (or average) is a measure of central tendency that represents the central value of a data set. It is one of the most fundamental concepts in data analysis, used across fields ranging from finance and economics to science and engineering. While the arithmetic mean is the most commonly used, there are several types of means—each with its own properties, assumptions, and appropriate use cases.
For a data set x₁, x₂, … , xn, the arithmetic mean is:
x̄ = (Σ xᵢ) / n
where Σ denotes summation and n is the number of observations.
The concept of the mean dates back to ancient civilizations. The Babylonians and Egyptians used averaging techniques for land measurement and taxation. The modern notion of the arithmetic mean was formalized by mathematicians such as Carl Friedrich Gauss and Adrien-Marie Legendre in the context of the method of least squares. Today, the mean is a cornerstone of descriptive and inferential statistics, forming the basis for more advanced techniques like regression analysis, ANOVA, and hypothesis testing.
Beyond the arithmetic mean, the geometric mean and harmonic mean serve specialized purposes. The geometric mean is used when dealing with rates of change, growth factors, and multiplicative processes. The harmonic mean is appropriate for rates and ratios, such as average speed or average price-earnings ratios. Choosing the correct mean is critical to obtaining meaningful insights from data.
| Type | Formula | Best Used For | Key Property |
|---|---|---|---|
| Arithmetic Mean | x̄ = (Σx)/n | General-purpose average; symmetric data without outliers | Sum of deviations from mean = 0 |
| Geometric Mean | G = (Πx)1/n | Growth rates, investment returns, multiplicative processes | Less sensitive to large values; always ≤ arithmetic mean |
| Harmonic Mean | H = n / Σ(1/x) | Rates, ratios, average speed over multiple segments | Smallest of the three; useful for inverse relationships |
| Median | Middle value (sorted) | Skewed distributions; robust to outliers | 50th percentile; divides data into two halves |
| Mode | Most frequent value | Categorical data; identifying peaks in distribution | Can be multiple modes (bimodal, multimodal) |
The relationships between these measures: For any positive data set, H ≤ G ≤ x̄. Equality holds only when all values are identical. This inequality (the AM-GM-HM inequality) is a fundamental result in mathematics.
An investor tracks annual returns over five years: +10%, -5%, +15%, +8%, +12%. The arithmetic mean return is 8.0% per year. However, due to compounding, the geometric mean (the true average growth rate) is approximately 7.6% per year. The harmonic mean is not applicable here. Using the arithmetic mean would overstate the actual compounded return. This illustrates why the geometric mean is the correct choice for growth rates and investment performance.
Key insight: Always consider the nature of your data before selecting a mean. The arithmetic mean is appropriate for additive processes, while the geometric mean is essential for multiplicative processes.
For any set of positive real numbers, the arithmetic mean is always greater than or equal to the geometric mean, which is always greater than or equal to the harmonic mean:
x̄ ≥ G ≥ H
Equality holds if and only if all numbers are equal. This inequality has far-reaching implications in optimization, economics, and information theory. It underpins concepts such as the Cauchy-Schwarz inequality, Jensen's inequality, and the Gibbs inequality in information theory.
The AM-GM inequality is a powerful tool for proving other mathematical results and is often used in competitive mathematics and advanced statistical theory.
A climate scientist analyzes daily temperatures over a year: temperatures range from -10°C in winter to 35°C in summer. The arithmetic mean temperature is 12°C. However, the geometric mean is not appropriate for temperature data (which can be negative). The median (13°C) provides a better sense of the "typical" temperature. For environmental data, understanding the distribution is key, and using multiple measures (mean, median, mode, variance) gives a complete picture.