Box Plot Maker

How the Box Plot Maker works

Box plot maker: what to paste

Two layouts work. The wide layout has a header row naming each group and one column of numbers beneath it, which is how most spreadsheets hold repeated measurements; columns can be different lengths and blank cells are ignored. The long layout has exactly two columns, a group name and a value, with one row per observation, which is how databases and survey exports usually come. Cells can be separated by commas or tabs, so a block copied from Excel, Google Sheets or Numbers pastes straight in.

How each box is calculated

Values in a group are sorted and the first quartile, median and third quartile are found by linear interpolation, the same method as Excel's QUARTILE.INC and R's default type 7. The box spans the interquartile range from Q1 to Q3, with a heavy line at the median. The whiskers use the Tukey rule: they reach to the smallest and largest values that lie within 1.5 times the interquartile range of the box, and anything beyond that is drawn as a hollow circle and listed as an outlier. Hover a box for its full five-number summary and the stats panel repeats the numbers for every group.

Options

Orientation switches between vertical boxes with groups along the bottom and horizontal boxes with groups down the side, which suits long group names. Mean marker adds a white dot at the arithmetic mean of each group, useful because a mean far from the median signals a skewed distribution. Data points overlays every observation as a small jittered dot, so the shape of the data behind each box is visible; it is worth switching on when groups are small, since a box drawn from five numbers implies more precision than it has. Value axis label names the measurement, for example "Response time (ms)".

Reading the chart

The median line shows the typical value and the box height shows how spread out the middle half is. Two boxes that do not overlap are clearly different; overlapping boxes may or may not be, and a formal test is needed to say. Whiskers of very different lengths, or a median close to one end of the box, indicate skew. Outliers are not errors by definition; they are simply values far from the rest and deserve a look at the source data before being excluded.

Exporting

Download SVG gives a vector file with editable text and one group per element. Download PNG renders at one, two or three times the preview width. Copy SVG puts the markup on the clipboard.

Related tools