# Histogram

**URL:** https://kaizumi.com/dictionary/histogram

**Description:** A histogram is a bar chart showing the frequency distribution of continuous data, revealing patterns in process variation and shape.

**Category:** six-sigma

**Tags:** metrics, visual-management, foundational

## Definition

A histogram is a graphical representation of data distribution, displaying how frequently values occur across different ranges. Data is grouped into bins (intervals), and the height of each bar represents the count of values falling within that range. Histograms reveal distribution shape (normal, skewed, bimodal), spread (variation), and centering relative to specifications. They are fundamental tools in Six Sigma for understanding process behavior and identifying problems invisible in simple statistics.

## Examples

### Manufacturing

> Shaft diameter data showed acceptable mean and standard deviation. A histogram revealed the distribution was bimodal—two peaks indicating two different process states. Investigation found that day and night shifts had different machine settings. The simple statistics masked a controllable source of variation.

### Healthcare

> Patient wait times averaged 15 minutes—within target. A histogram showed right-skewed distribution with a long tail extending to 60 minutes. While most patients waited briefly, a significant group had very long waits. The average masked serious inequity requiring attention.

### Administrative

> Processing times averaged 3 days as expected. A histogram revealed a bimodal distribution—most completed in 1-2 days, but a second cluster took 7-8 days. Investigation found that certain document types triggered a different approval path. Targeting the long path improved overall performance.

## Key Points

- Histograms show distribution shape, spread, and centering
- Can reveal problems hidden in averages: bimodality, skewness, outliers
- Number of bins affects interpretation—too few or too many obscures patterns
- Compare to specification limits to visualize capability

## Common Misconceptions

**Normal distribution is always the goal.** Many processes naturally produce non-normal distributions. Time-based data often skews right; count data follows different distributions entirely. The goal is understanding your data's actual behavior, not forcing normality.

**One histogram tells the whole story.** A single histogram captures data at one point in time. Process behavior changes. Regular histograms over time, combined with control charts, provide complete understanding.

## Related Tool

Where a histogram exposes variation in your data, a Yamazumi exposes variation in work content across operators — stack each step's time against takt with the [Yamazumi chart builder](/tools/yamazumi).

Source: https://kaizumi.com/dictionary/histogram
Licence: free to quote and cite with attribution to Kaizumi.