Child pages
  • Box Plot (5.1)

Contents

Context

Data | Charts

New: Box Plot

A Box Plot (also known as a box-and-whisker plot) is a classical tool used in Descriptive Statistics to analyze the distributions of numerical variables.

It shows:

  • the First Quartile Q1
  • the Median Q2
  • the Last Quartile Q3 
  • the Minimum, i.e. the lowest datum still within 1.5 of the Interquartile Range (IQR = Q3 −  Q1) of the first quartile
  • the Maximum, i.e. the highest datum still within 1.5 IQR of the last quartile
  • the first Notch, where N is the number of observations
  • the last Notch. Notches are useful in offering a rough guide to significance of difference of medians.
  • the Mean
  • Lower Suspect values that are within the Minimum and 
  • Upper Suspect values that are within the Maximum and 
  • Lower Extreme values that are inferior to
  • Upper Extreme values that are superior to 
Example

You select the variable to be analyzed in the Parameter window.

 

Upon clicking on Display, the box plots are generated. Hovering over each element returns its description and numerical value in the upper part of the window.

You can also zoom in vertically by selecting the y-range as illustrated below. Double-clicking on the graph returns the default view.

Clicking on a Suspect Point or Extreme Point (as indicated in the header) brings up a table with the corresponding data record.

This option is useful for comparing the conditional distributions of a variable given the states of a Selector Variable.

The graph below allows to compare the distributions of two variables, Smoothness (Mean) and Smoothness (Worst), given the two states of the variable Diagnosis (B and M).