A box plot that reads well: median, quartiles, whiskers and outliers for each category, with n on every box and the quartile method stated. The free version is a complete box plot. Pro adds points with jitter, mean and notches, violin, groups and small multiples, sorting by median, reference lines and a statistics table.
Box: the middle half of the values, from Q1 (25th percentile) to Q3 (75th). Line inside: the median. Whiskers: up to the last value still considered normal (by default within 1.5 × the box height). Dots beyond the whiskers: outliers. n: the number of rows behind the box.
If the median is not centred in the box, the values are skewed towards the longer side. A tall box means a spread-out, less predictable result; a short box, a consistent one. Compare medians to see which category is higher, and box heights to see which is more stable. With very few values (under about 10 per box), turn on all points (Pro) to see each one.
| Field | Kind | Required | What it does |
|---|---|---|---|
| Category | Grouping | Yes | One box per value: region, product, team. Add a hierarchy to drill down; with Expand all, one box per combination of levels. |
| Value | Measure | Yes, one field | The measure whose distribution is drawn, evaluated for each row of Detail. Rows with an empty or non-numeric value are not counted in n. |
| Detail (rows) | Grouping | Yes in practice, one field | The ID of each observation — order, employee, ticket, inspection. Without it Power BI sends one total per category and there is no distribution; the visual says so. |
| Group | Grouping | No — Pro, one field | Boxes side by side inside each category, coloured by group, with a legend. Without Pro the field is ignored and the free result is drawn. |
| Small multiples | Grouping | No — Pro, one field | One panel per value, all on the same value scale. |
| Tooltips | Measure | No | Extra measures in the tooltip, shown as the average over the rows of the box. |
| Capability | Free | Pro |
|---|---|---|
| Boxes, whiskers and outliers per category | ✓ | ✓ |
| Whiskers at 1.5 × IQR (Tukey) or minimum to maximum | ✓ | ✓ |
| Outliers always drawn, also on top of the points | ✓ | ✓ |
| n under each box, median label; n, min, quartiles, median, max, mean, outlier count and quartile method in the tooltip | ✓ | ✓ |
| Quartiles: inclusive method (Excel QUARTILE.INC) | ✓ | ✓ |
| Vertical or horizontal; 6 style templates, including IBCS style | ✓ | ✓ |
| Box colour with conditional formatting (fx) | ✓ | ✓ |
| Sort by data order or name | ✓ | ✓ |
| Axes: title, gridlines, range, font, size, colour, bold | ✓ | ✓ |
| Cross-filtering, bookmarks, drilldown, context menu, keyboard, high contrast | ✓ | ✓ |
| Format on the canvas, start screen, 12 languages | ✓ | ✓ |
| Group: side-by-side boxes with legend | — | ✓ |
| Small multiples, with columns, spacing, titles, panel background and border | — | ✓ |
| All points with jitter (separation, size, colour, border, opacity) | — | ✓ |
| Mean marker and notches (approximate 95 % interval of the median, McGill) | — | ✓ |
| Violin (density) behind each box | — | ✓ |
| Colour by quartile | — | ✓ |
| Exclusive quartiles (Excel QUARTILE.EXC) and Tukey hinges | — | ✓ |
| Whiskers at the 5th–95th or 10th–90th percentile | — | ✓ |
| Sort by median, mean, IQR or count | — | ✓ |
| Two reference lines with labels | — | ✓ |
| Statistics table: n, min, Q1, median, Q3, max, mean, std dev, outliers | — | ✓ |
Turn on a Pro capability without a licence while editing and it is drawn working under a "Pro preview" watermark that names it. Reading view shows the free result. Pro is sold per user on Microsoft AppSource.
| Card | Settings |
|---|---|
| Style template | Clean, Classic, Minimal, Contrast, Compact, IBCS style. A template only fills what you have not set yourself. |
| Boxes | Orientation; box colour with conditional formatting (fx) — the rule is evaluated per row, so base it on the category (a colour column or a measure that is constant within each category); colour by quartile and high contrast override it; border colour and width; box width %; rounded corners; fill opacity; median line colour; colour by quartile (Pro) with lower and upper half colours. |
| Whiskers | Reach: 1.5 × IQR (Tukey), minimum to maximum, 5th–95th or 10th–90th percentile (Pro); caps; colour. |
| Statistics | Quartile method (inclusive; exclusive and Tukey hinges are Pro); n under each box; median label; mean marker and colour (Pro); notches (Pro). |
| Points | Show all points (Pro); separation of the points %; size; colour; border colour and width; opacity; outlier size and colour. |
| Violin | Show violin (Pro), drawn for boxes with at least 3 values; opacity; colour (empty = the colour of the box or group). |
| Sort | Data order, name, or median, mean, IQR and count (Pro), computed on all the rows of each category so groups and panels share the order; descending. |
| Reference lines | Two lines (Pro), each with value and label; line colour. |
| Value axis | Show; title and title text; gridlines; automatic range or minimum and maximum; font; bold; text size and colour. |
| Category axis | Show; font; bold; text size and colour. |
| Legend | Shown with a Group field; position top or bottom; text size and colour. |
| Statistics table | Show (Pro), below the chart, up to 35 % of the height with its own scroll; text size and colour; header background and text colour; alternate row colour. |
| Small multiples | Columns (0 = automatic); space between panels; title size, colour, bold and alignment; panel background and border. |
| Background | Colour of the whole visual. |
Format on the canvas: in format mode, click the boxes, an axis, the legend or the background to format them where they are.
The box holds the middle 50 % of the values, from the first quartile (Q1) to the third (Q3). The line inside is the median. The whiskers reach the last value that is still considered normal, and the dots beyond them are outliers. n under the box is the number of rows behind it.
With 1.5 × IQR (Tukey), the default, each whisker goes to the last value within 1.5 times the height of the box (Q3 − Q1) from the box. Anything beyond is an outlier. Minimum to maximum draws the full range and no outliers. The percentile options (Pro) end the whiskers at the 5th/95th or 10th/90th percentile, which reads as "90 % of cases are between…".
There are several accepted ways to compute quartiles and they differ with few rows. The default is the inclusive method, the same as Excel QUARTILE.INC and PERCENTILE.INC. Pro adds Excel QUARTILE.EXC and Tukey hinges. The method in use is shown in every tooltip.
Add the ID of each observation to Detail (rows). Without it Power BI aggregates the value to one number per category and there is nothing to distribute. The visual shows a note when Detail is empty.
With one row per ID in Detail, both give the same number per row. Use the aggregation that makes sense if an ID can have several rows.
Power BI sends the visual up to 30,000 rows at a time. In the Power BI service the visual asks for further blocks automatically, up to 40 more (about 1.2 million rows). In Power BI Desktop the visual uses the first 30,000 rows and does not request more. In both cases a warning icon in the visual header says how many rows were used when some were left out. Statistics always use every row received. With all points shown, up to about 4,000 points per box are drawn; the violin's density is estimated from up to 2,000 values per box.
Bookmarks keep the boxes you select and the filters of the page. Like every Power BI visual, they do not keep formatting such as orientation or points. To switch between two looks, use two copies of the visual and show one or the other in each bookmark.
While you edit a report without a licence, Pro features you turn on are drawn working under a watermark that names them, and Power BI shows its licence notice. Reading view shows the free result, with a small note naming the Pro features it leaves out. Nothing is lost: the settings apply as soon as a licence is active. Where Power BI provides no licence information (for example when you are not signed in to Power BI Desktop), the free result is shown and there is no preview.
On the visual's Microsoft AppSource page, choose Buy now (not Install free). Licences are per user and managed by Microsoft; restart Power BI Desktop after the licence is assigned. See tcviz.com/licensing.
With the measure's number format from the model, dynamic format strings included, in the report's language. Statistics that are not whole numbers (mean, interpolated quartiles, standard deviation) show up to two decimals when the format has none.
No. The visual makes no network requests and stores nothing outside your report. The links on the start screen open only when you click them, through Power BI.
Write to support@tcviz.com or open a thread in GitHub Discussions.