How to make a violin plot
- Copy the measurement columns to compare (or one measurement column plus a group column) together with the header.
- Click the column header (A) in the data table and paste - the first line becomes the column name. You can also open a CSV or Excel file.
- Select the columns to plot. When using a group column, select only one data column.
- Press Draw and check the body shapes, the horizontal lines (mean, median, min, max) and the automatic interpretation.
What is a violin plot?
A violin plot draws the density of the data as a symmetric curve. The wider the body, the more data near that value. A box plot summarizes only the quartiles, while a violin shows the actual shape of the distribution - the number of peaks and the skewness.
How it is calculated
For each series a Gaussian kernel density (bandwidth: Scott's rule) is evaluated at 100 points between the minimum and maximum and scaled so the widest half-width is 0.25 (the same method as matplotlib violinplot). The vertical bar spans min to max, and the horizontal lines mark the mean, median, minimum and maximum.
How to read a violin plot
- A waist with two peaks may mean data from two different conditions (equipment, raw material, etc.) are mixed.
- A body stretched to one side indicates a skewed distribution; the farther apart the mean and median lines, the stronger the skew.
- The automatic interpretation reports the number of outliers (IQR rule) and, for several series, compares medians and standard deviations.
Frequently asked questions
Should I use a box plot or a violin plot?
Use a box plot for summary statistics (quartiles, outliers) and a violin plot for the distribution shape. Looking at both is the most informative.
How many data points do I need?
Two or more draw a plot, but 20 or more per series are recommended for a meaningful density curve. If all values are equal, only the lines are drawn.
Can I split the data by group?
Yes. Choose one data column and a group column to draw one violin per group value.
Is my data sent to a server?
No. Everything is processed inside your browser and nothing is transmitted or stored.
Other chart tools
Need process capability, control charts or DOE?
The Qstat desktop program offers about 250 analysis features: histograms and box plots plus process capability (Cp, Cpk), control charts, measurement system analysis (MSA), design of experiments (DOE), regression and machine learning.
Try Qstat free for 30 days