How to Make a Histogram in Excel

To make a histogram in Excel, place the raw numeric observations in one column, select the values, choose Insert and then Histogram, and review the bin width or number of bins. The bars should represent numeric intervals and the vertical axis should count observations.

A histogram answers a distribution question: where do individual measurements cluster, how widely do they vary, and are there gaps, tails, or unusual values? It is not a bar chart with a different style. The data structure is the decisive difference. A histogram starts with raw observations such as twenty response times. A bar chart can start with named categories and their already summarized totals.

This guide uses fictional website response times in milliseconds. The menus target recent desktop versions of Excel, including Microsoft 365. Microsoft changes labels across editions, so use its chart creation guide and chart-type reference when your ribbon differs.

Use raw observations, not a pre-counted table

Download the fictional response-time CSV, or copy the values below. Every row after the header represents one measured request.

Response time (ms)
182
205
218
221
229
235
241
244
249
253
258
263
269
276
281
289
301
318
347
392

The example is fictional and is for demonstration only. It contains twenty raw measurements, not a frequency table. Excel will group these values into intervals and count how many observations fall into each interval. Keep units consistent: do not mix milliseconds and seconds in the same column.

A table such as “0–200 ms: 1, 200–300 ms: 15, above 300 ms: 4” has already been summarized. Do not paste the three counts into a histogram tool as if they were three raw measurements. Use the original twenty observations, or visualize the summarized ranges with a bar chart while clearly labeling the categories.

Clean the measurement column

1. Use a single header. Name the measure and include the unit, such as Response time (ms).
2. Keep one observation per row. A histogram needs the individual values in order to calculate interval counts.
3. Remove text from numeric cells. Put the unit in the header or number format rather than typing “ms” after every value.
4. Investigate blanks and errors. Do not silently replace a missing measurement with zero. Zero may be a valid measurement and would change the distribution.
5. Preserve unusual values until reviewed. An outlier can be a real observation, a measurement problem, or a data-entry error. Deleting it without an explanation changes the story.

Create the histogram in Excel

1. Open the CSV and confirm that the header and all twenty values appear in one numeric column.
2. Select the header and observations, excluding notes, totals, blank separator rows, and unrelated columns.
3. Choose Insert, select the statistical chart options, and choose Histogram. If the option is not present in your edition, consult Microsoft's current chart-type documentation rather than substituting a standard column chart without calculating bins.
4. Give the chart a precise title, such as “Distribution of fictional response times.” State the unit in the title or horizontal axis.
5. Open the horizontal-axis formatting options and review the bin settings. Compare more than one sensible bin width or bin count before interpreting the shape.
6. Confirm that the vertical axis represents the number of observations. If you need percentages or density, calculate and label that transformation explicitly.

Choose bins deliberately

Bins divide a numeric scale into intervals. A bin from 200 up to 225 ms might contain 205, 218, and 221. Adjacent histogram bars touch because the intervals belong to one continuous scale. The exact boundary convention matters: software commonly includes one boundary and excludes the other so that a value is counted once.

Too few bins hide variation. The fictional values could collapse into a broad central bar and a broad upper bar, concealing the gradual tail. Too many bins create empty gaps and isolated spikes that make a small sample look more structured than it is. There is no single visual that removes this judgment. Compare several reasonable settings and report the one that communicates the distribution without manufacturing a pattern.

The sample size also limits the conclusion. Twenty observations can demonstrate a workflow, but they are not strong evidence of long-term performance. A histogram shows the collected sample. It does not prove that the next observation will follow the same shape.

Read the histogram carefully

Start by describing visible features rather than assigning a cause. In the fictional data, many observations are between roughly 220 and 290 ms, while a smaller number extend above 300 ms. That description may suggest a right tail. It does not identify the requests, devices, locations, or server conditions responsible for the slower values.

Look for a central cluster, overall spread, empty ranges, multiple clusters, and isolated intervals. Compare the graph with summary statistics and the source rows. The NIST Engineering Statistics Handbook's histogram overview is a useful reference for what a histogram can reveal and why its interpretation belongs to exploratory data analysis.

Common histogram mistakes

Create the same distribution in Charttoolbox

Use Excel when the histogram should stay inside the workbook. If you want a separate PNG or SVG from a small list of values, open the histogram maker. Charttoolbox reads CSV, TSV, pasted cells, and the first worksheet of XLS or XLSX files. It does not import an existing Excel chart or its formatting.

1. Upload or paste the raw observations. Review the detected Value field and confirm the unit.
2. Begin with Auto bins, then compare another bin count. The tool uses equal-width bins, with Auto based on a Sturges-style count capped at thirty.
3. Add a title, source, takeaway, or footnote, select a presentation style, and export PNG or SVG.

The tool accepts one numeric column of raw observations and up to fifty bins. Blank cells, text, infinities, and mixed percentage and ordinary-number formats must be corrected. Constant values are displayed in one padded interval rather than being turned into a fake spread. Spreadsheet contents are processed locally in the browser; optional analytics requires consent and does not receive the table.

Histogram or bar chart?

Choose a histogram when each row is an individual numeric observation and the question concerns distribution. Choose a bar chart when the bars represent named categories such as device type, country, or product. Histogram intervals have numeric order and normally touch. Bar categories can be reordered and normally have gaps.

Read the full histogram versus bar chart comparison if the source is already summarized. For change across dates, choose a line graph. For the relationship between two numeric measurements on the same observations, choose a scatter plot.