Gmisc-package: Collection of functions for plotting relations, generating...

In raredd/Gmisc0: Collection of functions for plotting relations, generating tables, and more.

Description

This is a collection of functions that I've found useful in my research. The package is inspired by Frank Harrell's Hmisc package. The main focus is on tables, plots, and knitr-integration.

Awesome tables

For tables you'll find the convenient htmlTable that I have used for advanced table layout. A major focus has been to have it compatible with LibreOffice (you can copy/past from there into word) as I generally want to be able to send my documents to a journal in .doc/.docx format. Note: it is now in RStudio possible to copy->paste directly from the viewer into a MS Word document with minimal layout loss.

The getDescriptionStatsBy is a straight forward function that aims at helping you to generate descriptive table stratified by different variables. In other words, the function returns everything you need for generating a Table 1 ready for publication. This function is accompanied by the describeMean, describeMedian, describeProp, and describeFactors functions.

Convenient knitr-helpers

One of the main priorities of this package is to make the preparation of publication-ready manuscripts through the knitr-package. The figCapNo can be used for automated figure counting. The pvalueFormatter tries to simplify rounding of p-values, e.g. you may be ok with just 0.0005 as a p-value but when you come close to the "magic" 0.05 value you may want to have two significant digits, i.e. 0.048 instead of just 0.05. The outputInt simply transforms a large integer digit to proper formatting.

Some fancy plots

The forest plot function, forestplot2, is a more general version of the original rmeta-packages forestplot implementation. The aim is at using forest plots for more than just meta-analyses.

The transition plot function, transitionPlot, is for descriptive purposes. It tries to illustrate the size of change between one state and the next, i.e. a transition. This is basically a graph of based upon table(var1, var2).

The Singular value decomposition is a common method for reducing the number of variables. Unfortunately this compression can reduce the interpretability of the model. The getSvdMostInfluential function tries to remedy that by identifying the most influential elements from the V-matrix.

The getTicks tries to format ticks for plots in a nicer way. The major use is for exponentials where ticks are generated using the 2^n since a doubling is a concept easy to grasp even for non-statisticians.

Other stuff

The insertRowAndKeepAttr simply adds a row while remembering all the attributes previously set by using the copyAllNewAttributes. The mergeLists tries to merge lists that do not have identical elements.

raredd/Gmisc0 documentation built on May 27, 2019, 2:02 a.m.