What Is a Histogram?
A histogram bins a continuous numeric measure into ranges and shows how many records fall into each range as a bar. Unlike a bar chart, which compares distinct categories, a histogram reveals the shape of a single measure's distribution — where values cluster, how spread out they are, and whether there are one or several peaks.
It answers a different question than most charts on a dashboard. A bar chart of sales by region tells you who sold the most. A histogram of order values tells you what a typical order actually looks like — and whether "typical" is even a meaningful word for that data.
Why the Average Alone Can Mislead
A single average hides everything about shape. Two datasets can share the exact same mean while looking completely different: one tightly clustered around it, the other bimodal with two distinct clusters far from any "average" customer. A histogram is the fastest way to catch that difference before it distorts a decision.
Reading Mean vs. Median
Most histograms overlay two reference lines: the mean (average) and the median (the middle value). The gap between them tells a story on its own:
- Mean ≈ Median — the distribution is roughly symmetric.
- Mean > Median — right-skewed: a small number of unusually high values are pulling the average upward (common in revenue, order size, response time).
- Mean < Median — left-skewed: a small number of unusually low values are pulling the average down.
When the gap is large, the average is a poor summary of "typical" — the median or the mode is more representative.
What to Look For
- Shape — one clean peak (unimodal) suggests a single underlying process. Two peaks (bimodal) usually means two distinct groups are being mixed together in the same measure.
- Spread — a tall narrow histogram means low variability; a wide flat one means high variability, even if the average is identical.
- Outliers and tails — a long tail on one side flags extreme values worth investigating separately, since they can distort both the average and any statistical model built on the data.
- Bin count — too few bins hides real structure; too many turns the shape into noise. A common rule of thumb is roughly √n bins for n data points, adjusted by eye.
Histogram Pro for Power BI
Histogram Pro bins a measure into a distribution and overlays mean/median reference lines automatically, with full control over bin count, width and color scales — the statistical reading described above, built into the visual.
What it adds
- Configurable bins from 2 to 100 (Pro) instead of a fixed 10 — zoom into fine detail or pull back to an executive-level summary.
- Live statistics panel: n, mean, median, standard deviation, min, max — six numbers next to the shape, no separate card visuals needed.
- Up to 5 reference lines (Pro): mean, median, Q1, Q3, and a custom target/benchmark.
- Outlier trimming (Pro) — exclude a top/bottom percentage before binning, without touching the underlying data model.
- Normal curve overlay — see how closely the actual bars follow a theoretical normal distribution, useful for validating whether statistical assumptions hold.
- IBCS monochrome mode for standards-based financial reporting.
Getting Started
- Drag a numeric field into Values — the histogram renders instantly with 10 bins, mean and median lines.
- Add a Category field (optional) to enable cross-filtering with the rest of the report.
- Tune the bin count in the Format Pane to trade off detail vs. readability.
- Enable the normal curve overlay to check for skew or bimodality.
- Add a benchmark line for your SLA or target value for immediate context.
If a single average is the only number your report shows for a measure, a histogram is usually the fastest way to find out what that average is hiding.