Box plot maker

Put each group (machine, shift, supplier) in its own column with the name in the first row, copy the columns and paste them into the values box. Columns may have different lengths. If your data are in a group column and a value column, choose the Stacked layout.

Vertical is the classic layout, good for up to 6 or 8 groups with short names. Horizontal works better with many groups, long names and on phones, because each group gets its own row. Use the Vertical and Horizontal buttons above the chart; the choice also applies to the PNG and the PDFs.

With the Minitab method: Q1, median and Q3 at positions (N + 1)·0.25, (N + 1)·0.5 and (N + 1)·0.75 of the sorted data, interpolating between neighbors. Excel’s QUARTILE.INC uses another rule and may give slightly different Q1 and Q3 with small samples.

A value more than 1.5 × IQR below Q1 or above Q3. The page shows each one with its value and spreadsheet row, so you can go straight to the record. An outlier calls for investigation (measurement error, typing, a part from another lot); only remove it once the cause is found.

The notch shows the 95% confidence interval of the median. When the notches of two groups do not overlap, the difference between the medians is probably real. For a formal conclusion, use the Kruskal-Wallis test or ANOVA with the button below.

A mean (◆) far from the median signals skewness or outliers pulling the value. The points show how much data each box holds and whether there are clusters the box hides; with more than 50 values per group they are hidden in automatic mode.

No. The box plot shows where the differences are and whether the spreads are similar, which is an ANOVA condition. If the boxes look different, confirm with ANOVA (means) or Kruskal-Wallis (medians); if one box is much longer, investigate that group first.