If an image is worth than thousand words: how much a box plot can say?
- Dennis Denis Ávila
- Víctor Manuel Ramírez-Arrieta
DOIAbstract
Graphic data visualization is essential to Exploratory Data Analysis and precedes any statistical analysis. Box plots are one of the most widely used graphical representation to show descriptive statistics of data in a sample and to represent multiple comparisons. This graphic, however, if careless used can markedly distort data interpretations, and for this reason in this paper we describe box plot pitfalls and show several ways to improve them using R programming flexibility. By adding jittered raw data, histograms and density curves graphic variants can be produced, which had been described in recent literature with other names such as violin plots, pirate plots, raincloud plots, beanplots, sinaplots, among others. We also present the ap...
Add to ORKG