Boxplot Creator

Box Plot Maker

Paste one list of numbers, or several columns to compare groups side by side. Get box plots, the five-number summary, outliers, mean and standard deviation, with a choice of quartile method. Download the chart as a PNG. Everything runs in your browser and nothing is uploaded.

One group: just numbers, separated by commas, spaces or new lines. Several groups: a header row with one column per group (copy from Excel or Sheets), or two columns of group name and value.

Box = Q1 to Q3, line = median, diamond = mean, whiskers = furthest points inside the fences, hollow circles = outliers.

    What a box plot shows

    A box plot (box-and-whisker plot) summarises a distribution with five numbers: the minimum, first quartile (Q1), median, third quartile (Q3) and maximum. The box covers the middle half of the data, the line inside it marks the median, and the whiskers show how far the rest of the data extends. John Tukey popularised the plot in his 1977 book Exploratory Data Analysis.

    Because it uses medians and quartiles rather than means and standard deviations, a box plot is not thrown off by a few extreme values and makes no assumption that the data are normally distributed. That makes it a good first look at skewed data such as incomes, waiting times or test scores, and an efficient way to compare several groups on one axis.

    How to read one

    • Median position. If the median line sits near the middle of the box, the middle half of the data is roughly symmetric. If it sits close to Q1, the data bunch at the low end and trail off high (right skew), and the reverse for left skew.
    • Box length (IQR). A longer box means more variation in the typical values. The IQR is Q3 − Q1.
    • Whisker lengths. A long whisker on one side means a stretched tail on that side.
    • Outliers. Points beyond the fences Q1 − 1.5×IQR and Q3 + 1.5×IQR are drawn separately. They are unusual relative to the middle of the data, which is not the same as being mistakes.
    • Mean versus median. A mean well above the median points to a long upper tail, and a mean below the median points to a long lower tail.
    • Comparing groups. Compare medians first, then box lengths. Boxes that barely overlap suggest a clear difference; heavily overlapping boxes suggest the groups are similar. This is an informal impression, not a statistical test.

    Why different tools give different quartiles

    There is no single definition of a quartile for a finite sample. Hyndman and Fan (1996) catalogue nine in common use. This tool offers the three that matter most in practice:

    • Inclusive (type 7). Interpolates at position (n−1)×p. It is the default in R, Excel’s QUARTILE and QUARTILE.INC, Python’s NumPy and pandas, and Google Sheets.
    • Exclusive (type 6). Interpolates at position (n+1)×p. It matches Excel’s QUARTILE.EXC and gives slightly wider quartiles on small samples.
    • Median of halves. Split the sorted data at the median (leaving the median out when n is odd) and take the median of each half. This is the method many textbooks teach.

    For large samples the three agree closely. For small ones they can differ visibly, which is why a hand calculation sometimes does not match software. State the method you used when you report results.

    Worked example you can check by hand

    Take the nine values 2, 4, 5, 6, 7, 8, 9, 10, 30. The median is 7 and the mean is 81/9 = 9.

    Inclusive (type 7): Q1 = 5, Q3 = 9, IQR = 4. The fences are −1 and 15, so 30 is an outlier and the upper whisker stops at 10.
    Exclusive (type 6) and median of halves: Q1 = 4.5, Q3 = 9.5, IQR = 5. The fences are −3 and 17, and 30 is still an outlier.

    The mean (9) is above the median (7), and the single outlier explains why. Paste these numbers into the tool and switch the quartile method to see all three results.

    Mistakes that make box plots misleading

    • Treating every outlier as an error. The 1.5×IQR rule is a screening convention. Check the source of each flagged value before removing it, and report what you did.
    • Plotting very few points. With fewer than about 8 values per group, quartiles are unstable. Tick “Show all points” to see the real data.
    • Hiding the shape. A box plot cannot show two peaks. A bimodal dataset can look identical to a unimodal one, so also look at a histogram.
    • Ignoring unequal group sizes. A box built from 6 values is far less reliable than one built from 600, but they look the same. The n row in the table matters.
    • Mixing quartile methods. Comparing numbers from tools that use different definitions creates differences that are not in the data.
    • Reading the box as a confidence interval. The box describes the spread of the data, not the uncertainty of the median.

    Box plot, histogram or violin plot?

    ChartBest forWeakness
    Box plotComparing several groups compactly; spotting skew and outliersHides multiple peaks and exact shape
    HistogramSeeing the full shape of one distributionShape depends on bin width; hard to compare many groups
    Violin plotShape and comparison togetherLess familiar; needs enough data for a smooth density

    Frequently asked questions

    What input formats are accepted?

    A plain list of numbers separated by commas, spaces, semicolons or line breaks; a table with a header row where each numeric column becomes a group; or two columns where one holds group names (up to eight different names) and the other holds numbers. Empty cells are ignored, so groups can have different lengths.

    Which quartile method should I choose?

    Use the one your course, journal or software expects. If there is no requirement, the default inclusive method matches Excel, R, pandas and Google Sheets, so your results will agree with the tools most people use.

    What does the 3 × IQR option do?

    It moves the fences further out so only extreme outliers are flagged. Some analysts call values between 1.5 and 3 IQR “mild” and beyond 3 IQR “extreme”.

    Why is there no outlier on my plot when one value looks far away?

    If the box is long, the fences are far away too. A value only counts as an outlier relative to the spread of the middle half. With very small samples the fences can also be wide.

    Should I remove outliers?

    Only with a reason, such as a recording error or a case that does not belong to the population you study. Otherwise keep them, and consider reporting results with and without them.

    Is my data uploaded anywhere?

    No. The statistics and drawing run in JavaScript on this page. You can confirm in the Network tab of your browser’s developer tools that no data is sent.

    Further reading

    • Tukey, J. W. (1977). Exploratory Data Analysis. Addison-Wesley.
    • Hyndman, R. J., & Fan, Y. (1996). Sample quantiles in statistical packages. The American Statistician, 50(4), 361–365.
    • Frigge, M., Hoaglin, D. C., & Iglewicz, B. (1989). Some implementations of the boxplot. The American Statistician, 43(1), 50–54.

    Box Plot Maker by Xcelario, part of the Academic Tools collection. Statistics follow the definitions above; with the default settings quartiles match Excel’s QUARTILE.INC and R’s quantile type 7. This tool supports learning and exploratory analysis; for publication or high-stakes decisions, confirm results in dedicated statistical software. Last updated November 2026.