Statistics › 📊 Describing Data

Histograms & the Shape of Data

How binning turns a pile of numbers into a picture — and how that picture's shape tells you where the center really is.

Intro StatisticsAP Statistics level

💡

The big idea: A histogram takes a long list of numbers, slices the number line into equal-width bins, and stacks up how many values fall in each one. The resulting silhouette — symmetric, skewed, flat, or two-humped — is the fastest way to see what a data set is really like. And once you can read that shape, you can predict something subtle: when the picture leans to one side, the mean gets dragged toward the long tail while the median stays near the crowd.

🎯 By the end, you'll be able to

Explain how a histogram is built from bins and frequencies
Predict how changing the bin width makes the same data look smoother or spikier
Name a distribution's shape: symmetric, right- or left-skewed, uniform, or bimodal
Describe how skew pulls the mean away from the median, and use that gap to read a distribution's direction

📎 You should already know

Mean, median, and mode
Reading a basic bar chart

From a list of numbers to a picture

Suppose someone hands you the wait times of 500 customers, or the scores on a class quiz. As a raw list it is unreadable — just a wall of numbers. A histogram fixes that with one simple trick.

Chop the number line into equal-width slices called bins, then count how many values fall into each slice and draw a bar that tall. A bin holding lots of values gets a tall bar; an empty stretch gets no bar at all. The height of each bar is just a frequency — a count. Stand back, and those bars trace out the shape of the data.

🔑 What a histogram actually shows

A histogram is not the same as a bar chart. A bar chart compares separate categories; a histogram takes one numeric variable, splits its range into touching bins of equal width, and the bar height is the number of values (the frequency) in each bin. Bars touch because the bins are adjacent slices of a continuous line.

Bin width changes the picture

Here is the catch that trips people up: the same data can look very different depending on how many bins you use. Bin width is a decision you make, not something baked into the numbers.

Too many, very narrow bins: each bin catches only a value or two, so the histogram looks spiky and full of gaps — you see random noise, not the pattern.
Too few, very wide bins: everything is crushed into two or three fat bars, hiding features like a second peak.

The number of bins and the bin width are two sides of one coin — for a fixed range, more bins means each is narrower.

\[ \text{bin width} = \frac{\text{max} - \text{min}}{\text{number of bins}} \]

For a fixed data range, using more bins makes each one narrower. There is no single correct choice — you tune it until the real shape shows through.

Try it: change the bins, change the shape

In the tool below, start with the Bins slider. Drag it high and the bars turn jagged and gappy; drag it low and the shape smooths into a few blocky bars — yet the underlying values never changed. Then use the Population menu to switch shapes, and watch the two dashed markers: the blue mean line and the green median line. On the right-skewed data they pull apart; on the bell-shaped data they sit almost on top of each other. Resample draws a fresh batch so you can see the shape is a property of the population, not one lucky sample.

🎮 Interactive: histogram shape, bins, mean vs median LIVE

Pick a population shape, then slide the number of bins to see the same data look smooth or spiky. The blue dashed line is the mean, the green dashed line is the median — notice how skew separates them while symmetry keeps them together.

✨ A short vocabulary of shape

Symmetric: the left and right halves are near mirror images (a bell curve is the classic case). Right-skewed: a long thin tail stretches to the right, with the bulk of the data on the left (incomes, wait times, house prices). Left-skewed: the long tail points left. Uniform: every bin is about the same height — a flat rectangle. Bimodal: two separate humps, which usually means two different groups are mixed together.

How skew moves the mean and the median

Shape is not just decoration — it tells you which summary of the center to trust. The mean is an average, so a few extreme values in a long tail tug it in their direction. The median is just the middle value once the data is sorted, so it barely notices those outliers.

Put those two facts together and you get a reliable rule of thumb: in a right-skewed distribution the tail on the right hauls the mean above the median; in a left-skewed one the mean sits below the median; and in a symmetric distribution they land in essentially the same place.

⚠️ Which center should you report?

When data is strongly skewed or has outliers, the mean can be misleading — a handful of huge values inflates it above what a typical case looks like. That is exactly why incomes and home prices are usually reported as a median: it stays close to the crowd. Read the shape first, then choose your summary.

📝 Worked example: A small shop records five repair times (in hours): 3, 4, 4, 5, 34. Find the mean and the median, and use them to describe the shape.

Mean: add them up and divide by 5. Sum = 3 + 4 + 4 + 5 + 34 = 50, so the mean = 50 ÷ 5 = 10 hours.
Median: sort the values (they already are: 3, 4, 4, 5, 34) and take the middle one. With five values the middle is the 3rd, so the median = 4 hours.
Compare: the mean (10) is far above the median (4). The single 34-hour job is a long right tail that drags the mean upward while the median stays near the typical case.

✓ Mean = 10 h, median = 4 h. Because the mean sits well above the median, the data is right-skewed — and the median (4 h) better describes a typical repair than the mean does.

Check your understanding

1. In a histogram, what does the height of each bar represent?

A histogram bar's height is a count: how many data values landed in that bin's slice of the number line.

2. You split the same 40 data points into far too many narrow bins. The histogram will most likely look…

With too many bins, each catches only a value or two, so the picture turns jagged and gappy and shows noise instead of the underlying shape.

3. A distribution is right-skewed, with a long tail stretching to the right. Which is usually true?

The long right tail pulls the mean toward the large values, so in right-skewed data the mean is typically greater than the median.

4. For the data set 2, 3, 3, 4, 50, the mean is 12.4 and the median is 3. This tells you the data is…

The mean (12.4) is far above the median (3) because the value 50 forms a long right tail — the signature of right skew.

✅ Key takeaways

A histogram slices the number line into equal-width bins and shows how many values land in each — the bar heights are frequencies (counts).
Bin width is a choice, not a fact: too many bins looks jagged and full of gaps, too few hides the shape. Adjust it until the real pattern is clear.
Learn the vocabulary of shape — symmetric, right-skewed, left-skewed, uniform, and bimodal — because the shape tells you which summary numbers to trust.
In a symmetric distribution the mean and median sit together; skew separates them.
The mean is pulled toward the long tail, so mean > median signals right skew and mean < median signals left skew. The median resists outliers.