Histogram

A histogram is a chart that shows the distribution of a single numerical variable by dividing its range into intervals, called bins, and drawing a bar for each bin whose height represents how many observations fall within it. It's the standard first visualization reached for when exploring a new numerical variable, because it directly answers the question a table of summary statistics alone cannot: not just where the center and spread of the data are, but what overall shape the distribution takes — symmetric, skewed, bimodal, or containing unusual gaps or spikes that a mean and standard deviation would never reveal on their own.

The choice of bin width is not a cosmetic decision but a substantive analytical one, because it directly controls how much of the underlying data's real structure the histogram actually reveals. Too few, overly wide bins smooth over real structure — a bimodal distribution (two distinct underlying groups mixed together) can collapse into what looks like a single smooth hump if the bins are too wide to separate the two peaks. Too many, overly narrow bins introduce noisy, jagged structure that doesn't represent anything statistically meaningful, just random sampling variation in which exact bin each nearby data point happened to fall into. Several formal rules exist for choosing a reasonable default bin width or count (Sturges' rule, the Freedman-Diaconis rule, among others), but none is universally correct for every dataset, which is why experienced analysts typically try several different bin widths on the same data rather than trusting a single default setting's output uncritically.

A histogram is frequently confused with a bar chart because both use rectangular bars, but the two represent fundamentally different kinds of data and that difference has real visual consequences: a bar chart represents discrete categories, which have no inherent order or continuity between them (and are conventionally drawn with gaps between bars to reflect that discreteness), while a histogram represents a continuous numerical variable divided into adjacent, ordered bins (conventionally drawn with no gaps between bars, since the bins represent continuous, adjoining ranges of the same underlying variable, not separate unrelated categories). This distinction isn't merely stylistic — reordering a bar chart's categories doesn't change its meaning, while reordering a histogram's bins would be nonsensical, since the bins exist in an inherent numerical sequence that a bar chart's categories don't share.

A histogram directly visualizes several of the key descriptive statistics concepts it's typically paired with in analysis: the distribution's modal bin (tallest bar) corresponds to the mode, its overall symmetry or lack thereof directly shows skewness, and the presence of isolated bars far from the main cluster of data visually flags potential outliers — making a histogram a fast, intuitive way to sanity-check the numerical summary statistics an analysis has already computed, catching a case where the numbers alone look plausible but the actual shape of the distribution reveals something the summary statistics didn't make obvious.

Read a histogram as the standard way to see a single numerical variable's shape rather than just its summary numbers: its bin-width choice is a real analytical decision with documented formal guidance (Sturges' rule, Freedman-Diaconis) but no universally correct single answer, its visual distinction from a bar chart reflects a genuine underlying difference between continuous and categorical data rather than a stylistic convention, and its direct visual connection to skewness, modality, and outliers is exactly why it remains the first chart most analysts reach for when getting acquainted with a new numerical dataset.