The shape, spread, and location of a data set. Knowing the distribution of a data set can tell you a great deal about the data itself, and can be critical in selecting appropriate analyses and interpreting their results. You can assess a distribution through graphs, descriptive statistics, or more formally with a distribution identification tool.
Graphs like histograms can give instant insight into the distribution of a data set. Is the data clustered about a single area or are their multiple peaks or modes? Is it spread thinly over a broad range or tightly packed? Is the data skewed or symmetrical?
Descriptive statistics that describe the central tendency (mean, median) and spread (variance, standard deviation) of data with numeric values can be used to make comparisons with other data sets.
Finally, some "perfect" distributions can be defined on predictable graphic or mathematical terms and are referred to by name, like the Normal, Weibull, and Exponential distributions. The Normal distribution, for example, is always bell shaped and centered on a single mean value. Consistent percentages of values from a Normal distribution are contained within any number of standard deviations from the mean; for example, 68.26% of observations will always fall within one standard deviation from the mean. In practice, data only approximates these perfect distributions; if there is a close fit we say the data is "well modelled" by a given distribution. Minitab's Individual Distribution Identification tool can help you find the distribution that best fits your data.
Many analyses require data that is distributed in a particular way and may produce inaccurate results if anything else is used. Knowing that your data follows a particular distribution can alert you to discrepancies in your data. For example, right skewed data that is ordinarily well modeled by the normal distribution might signify that something has gone wrong in your process.