tukeyedar package houses functions used in Exploratory Data Analysis (EDA). Most functions are inspired by work published by John Tukey, David Hoaglin and Frederick Mosteller (see references at the bottom of this document). Note that this package is in beta mode, so use at your own discretion. Many of the plots generated from these functions are not necessarily geared for publication but are designed to focus the viewer’s attention on the patterns generated by the plots (hence the reason for light colored axes and missing axes labels for some of the plots ).
The functions available in this package include:
||Parallel boxplots with level and spread equalization|
||Trim lower values of a vector|
||Trim upper values of a vector|
||Trim lower records of a dataframe|
||Trim upper records of a dataframe|
||Re-express using Tukey powers or Box-Cox transformation|
||Letter value summaries|
||Generate scatter plot along with regression line and LOESS curve|
||Generate 3-point summary of data and plot half-slopes|
||Generate matrix of re-expressed univariate values based on ladder of powers|
||Generate matrix of re-expressed bivariate values and plot 3-point summary half-slopes|
||Fit a three-group resistant line to bivariate data|
This package can be installed from github (the installation process makes use of the
Note that the vignettes will not be automatically generated with the above command; note too that the vignettes are available on this website (see next section). If you want a local version of the vignettes, add the
build_vignettes = TRUE parameter.
The vignette will require that
dplyr be installed since the
eda_sl function relies on it. If
dplyr is not alreay installed, the aforementioned syntax will automatically install it for you.
If for some reason the vignettes are not created, you might want to reinstall the package with the
It’s strongly recommended that you read the vignettes. These can be accessed from this website:
If you chose to have the vignettes locally created when you installed the package then you can view them locally via
vignette("Introduction", package = "tukeyedar") and
vignette("RLine", package = "tukeyedar"). If you use a dark themed IDE, the vignettes may not render very well so you might opt to view them in a web browser via the functions
RShowDoc("Introduction", package = "tukeyedar") and
RShowDoc("RLine", package = "tukeyedar").
All functions start with
eda_. For example, to generate a three point summary plot of the
disp from the
mtcars dataset, type:
Note that most functions are pipe friendly. For example the following will work: