Analysis in R:I can see the relationship between indicators! The “corragram” package

Methods to show the relationship between indicators include scatter plots and correlation coefficients, and the corragram package allows you to output all of these methods together.

＜おすすめのRに関する書籍です＞

StanとRでベイズ統計モデリング Wonderful R 2 | 松浦健太郎, 石田基広

Stanによるベイズ統計モデリングの実践書。現実のデータ解析を念頭に、モデリングの考え方から丁寧に解説する。

Install Package

It can be introduced by running the following code.
URL: https://github.com/kwstat/corrgram

　install.packages("corrgram", type = "source")

#or

　library(devtools) #If devtools is not present, run right：install.packages("devtools")
  install_github("kwstat/corrgram")

Execute command

Here are two examples.
Note that the manual says you can choose between Pearson, Spearman, and Kendall for the correlation coefficient, but the correlation coefficient calculated in the panel.conf option is Pearson. Please note this!

#Reading data provided with the corrgram package, 74 rows and 14 columns.
data(auto)

#Example 1
#Default plot. Correlation coefficients can be selected from pearson, spearman, kendall.
corrgram(auto[, -c(1:2)], cor.method = "kendall")

#Example 2
#Plot with iris data supplied with R. Densities, scatter plots, and correlation coefficients for each variable can be plotted simultaneously.
#Note that the correlation coefficients in panel.conf are output in pearson even if the method is specified in cor.method. Please be aware of this!
corrgram(iris, lower.panel = panel.pts, upper.panel = panel.conf,
         diag.panel = panel.density, cor.method = "spearman")

Output Examples

Example 1

Example 2

少しでも、ウェブや実験の解析が楽になりますように！！

＜おすすめのRに関する書籍です＞

データ解析のための統計モデリング入門――一般化線形モデル・階層ベイズモデル・MCMC (確率と情報の科学)

一般化線形モデルから階層ベイズ・MCMCまでを一本の筋道で解説した、日本語の統計モデリング本の定番。

www.amazon.co.jp