Histogram

A histogram is a chart that shows how a single set of numbers is distributed. You take a numeric variable, say the dollar value of every order, and divide its range into consecutive intervals called bins. Then you count how many values fall into each bin and draw a bar for each one, with the bar's height representing that count. The result is a picture of the data's shape: where values cluster, how spread out they are, whether they lean to one side, and whether there are outliers or gaps. A histogram answers questions a single average cannot. Two datasets can share the same mean while looking completely different, one tightly clustered, the other spread wide or split into two humps, and the histogram makes that difference visible at a glance. It is one of the first tools any analyst reaches for when meeting a new numeric variable.

Reading a histogram is about reading its shape. A symmetric, bell-like shape suggests values gather around a central point and taper evenly on both sides. A shape with a long tail to the right, where a few large values stretch the picture, is right-skewed, common for things like income or order size where most values are modest and a few are very large. Two separate humps hint that the data mixes two different groups. The width of the bins matters, too: bins that are too wide hide detail and smooth the shape into a shapeless block, while bins that are too narrow make the chart jagged and noisy. Choosing a sensible number of bins is part of using a histogram well, because the same data can look smooth or spiky depending on how finely it is sliced.

A histogram is often confused with a bar chart, and the difference is more than cosmetic. A bar chart compares distinct categories, such as sales by region or revenue by product, where each bar is a separate group and the order of the bars is arbitrary. You could reorder the regions without changing the meaning, and there is usually a gap between the bars because the categories are unrelated. A histogram is different in kind: it shows the distribution of one continuous numeric variable, and its bars represent adjacent ranges of that variable, not separate categories. The bars sit next to each other with no gaps, because the ranges are continuous and touch, and their order is fixed because it follows the number line. Reorder the bins and the chart becomes meaningless.

The clearest way to hold them apart is to ask what the horizontal axis represents. In a bar chart, the axis is categorical, a list of names or labels with no inherent order. In a histogram, the axis is numeric and continuous, sliced into bins along a scale. That single distinction drives everything else: the gaps, the ordering, and what the chart is for. A bar chart answers which category is biggest; a histogram answers how a quantity is distributed. Using the wrong one misleads. Plotting continuous data as a bar chart with arbitrary buckets loses the sense of a distribution, and drawing categorical data as a touching-bar histogram falsely implies the categories form a continuous scale. Matching the chart to the data is the difference between clarity and confusion.

Using a histogram well starts with choosing bins thoughtfully. Too few bins flatten the data into a featureless block that hides the shape; too many make it jagged and read like noise. There is no single right number, so it pays to try a few and see which reveals the genuine structure without inventing detail that is not there. Look at the whole shape, not just the peak: note the center, the spread, any skew, and whether the tails or gaps suggest outliers or a mix of groups. Use the histogram to complement summary numbers rather than replace them, because the mean and the picture together tell more than either alone. And be honest about scale, since a histogram can be made to exaggerate or flatten differences by stretching or compressing its axes.

The failures come from careless bins and misreading. Bins that are far too wide or too narrow distort the shape and lead to wrong conclusions about how the data is really distributed. Confusing a histogram with a bar chart, treating continuous ranges as if they were separate categories, muddles what the chart is saying. Reading only the tallest bar and ignoring the tails misses outliers and skew that matter for decisions. And drawing a histogram from too little data produces a shape that is mostly noise, tempting an analyst to see patterns that would vanish with more observations. The discipline is to pick sensible bins, read the full shape, pair it with summary statistics, and remember that a histogram describes a distribution, which is exactly what a bar chart does not do.

Histogram — a chart that bins a numeric variable into ranges and plots each bin's count — reveals a distribution's shape, distinguishing it from the categorical bar chart.

Etymology: source.

Search interest for this term over the last five years:

View interest-over-time on Google Trends →

Areas of marketing where histogram is a core concern: