XLSTAT: Violin plot in Excel

 

Background

A violin plot is a method of plotting numerical data combining the vision of the probability density of each variable with its associated boxplot or dots.

The violin plot function developed in XLSTAT-R calls the geom_violin function from the ggplot2 package  in R (Wickham H).

Dataset for Example

The attached data correspond to the Iris dataset in the datasets package of R. This famous (Fisher’s or Anderson’s iris data set gives the measurements in centimeters of the variables sepal length and width and petal length, respectively, for 50 flowers from each of 3 species of iris. The species are Iris setosa, versicolor, and virginica . It was originally published by Fisher, R. A. (1936) The use of multiple measurements in taxonomic problems. Annals of Eugenics, 7, Part II, 179-188.

The  goal here is to compare the distributions of each variable (sepal length, sepal width, and petal length) according to the different species.

 

Setting Up

• Open XLSTAT.

• Click on the XLSTAT-R / plots menu and select the Violin plots (geom_violin) feature.

 

• In the Violin Plots dialog box, General tab:

  • Data: Select the range A1:C151.

  • Subsamples: Select the range D1:D151 (a qualitative variable to group violins by category).

  • Width: Adjust to control the spacing between violin plots within each category.

In the Options tab:

• Check Box plots to add boxplots to the violins, or Dot plots to add individual data points.

• Enable Trim to restrict violin tails to the actual data range.

In the Colors tab:

You can set Fill colors and Border colors using either a color palette (one color per group) or a single color for all violin plots.

• Click on the OK button to launch the computations.

• The results are displayed on a new Excel sheet. 

 

Interpreting the Results

The produced violin plot is displayed on the following graph:

Setosa appears to have smaller Sepal Length but similar Sepal Width compared to other species. Petal Length is clearly smaller for Setosa. These observations can be confirmed using hypothesis tests like ANOVA or t-tests.