What is a histogram?
A histogram groups measurements into equal-width ranges (bins) and shows how many values fall into each. Unlike a list of numbers or an average, it shows the distribution: where the process is centred, how wide it spreads and whether the shape hides something unusual. It is one of the seven basic quality tools and a quick first check before any capability study.
Common histogram shapes and what they mean
| Shape | Possible meaning |
|---|---|
| Bell-shaped, symmetric | Stable process with common-cause variation |
| Skewed | A natural boundary (zero, a stop), or a process that drifts in one direction |
| Two peaks (bimodal) | Two sources mixed: machines, shifts, lots. Stratify the data |
| Flat (plateau) | Several processes with different means combined, or a slowly drifting process |
| Cut-off edge | 100% inspection or sorting has removed values beyond a limit |
| Isolated island | A special cause, measurement error or a different population |
Histogram and process capability
With specification limits, the histogram shows whether the process fits. The tool calculates:
- Pp = (USL − LSL) / 6s: could the spread fit inside the specification if centred?
- Ppk = min(USL − mean, mean − LSL) / 3s: does it fit, given where it is centred?
- Expected out of specification, from a normal model, in % and ppm.
Capability numbers are only meaningful if the process is stable. Check stability first with a control chart.
Histogram example
The example shows 125 bottle fill weights against a specification of 497 to 503 g and a target of 500 g. The distribution is roughly symmetric, but it is too wide for the specification: Pp is only 0.86 and Ppk 0.71, three bottles are already outside the limits and the normal model predicts about 1.8% out of specification. The mean also sits about 0.5 g above target. Re-centring alone would not make the filler capable; the priority is to reduce variation, for example by stratifying the data by filling head.
How to use this Histogram tool
- Paste your measurements. Copy a column of values from Excel or your measurement system and paste it into the data box.
- Add specification limits. Enter the LSL, USL and target, if you have them, to see the data against the requirement.
- Check the bins. Keep the automatic bin width or set the number of bins or the bin width yourself.
- Read the shape. Look at centring, spread and shape. The tool flags skewness, off-centre processes and out-of-spec values.
- Review capability. Check Pp and Ppk and the expected percentage outside the specification.
- Export. Export the histogram and statistics to PDF, PNG or Excel.
Frequently asked questions
What is a histogram in quality control?
A histogram is a bar chart of how often measured values fall into consecutive ranges (bins). It shows the shape, centre and spread of a process, and how the data sits relative to the specification limits.
How many bins should a histogram have?
Too few bins hide the shape; too many make it noisy. A common guide is 5 to 20 bins, or about the square root of the number of values. This tool picks a bin width automatically (Freedman-Diaconis rule) and lets you set the number of bins or the bin width yourself.
What do histogram shapes tell me?
A symmetric bell shape suggests a stable process with common-cause variation. A skewed shape can come from a natural limit (such as zero). Two peaks often mean two machines, operators or materials mixed together, a sign to stratify. A cut-off edge can mean sorting or inspection removed out-of-spec parts.
What are Pp and Ppk?
Pp compares the specification width with the overall spread of the data: (USL - LSL) / 6s. Ppk also accounts for centring: the distance from the mean to the nearest limit divided by 3s. A Ppk of 1.33 or more is a common target. Because the histogram uses the overall standard deviation, these are performance (Pp/Ppk) rather than capability (Cp/Cpk) indices.
How much data do I need for a histogram?
At least 50 values are recommended to see the shape reliably, and 100 or more is better. With fewer values the bars are strongly influenced by chance.
Can I paste data from Excel?
Yes. Copy a column of values in Excel or Google Sheets and paste it into the data box. Values separated by spaces, tabs, semicolons or new lines are accepted. You can export the histogram to PDF, PNG or Excel for free.