R: A visualization of a correlation matrix.

corrplot {corrplot}

R Documentation

A visualization of a correlation matrix.

Description

A graphical display of a correlation matrix, confidence interval. The details are paid great attention to.

It can also visualize a general matrix by setting is.corr = FALSE.

Usage

corrplot(corr,
	method = c("circle", "square", "ellipse", "number", "shade", "color", "pie"),
	type = c("full", "lower", "upper"), add = FALSE,
	col = NULL, bg = "white", title = "",  is.corr = TRUE,		
	diag = TRUE, outline = FALSE, mar = c(0,0,0,0),
	addgrid.col = NULL, addCoef.col = NULL, addCoefasPercent = FALSE, 

	order = c("original", "AOE", "FPC", "hclust", "alphabet"),
	hclust.method = c("complete", "ward", "single", "average",
                      "mcquitty", "median", "centroid"),
	addrect = NULL, rect.col = "black", rect.lwd = 2,

	tl.pos = NULL, tl.cex = 1,
	tl.col = "red", tl.offset = 0.4, tl.srt = 90,

	cl.pos = NULL, cl.lim = NULL,
	cl.length = NULL, cl.cex = 0.8, cl.ratio = 0.15, 
	cl.align.text = "c",cl.offset = 0.5,

	addshade = c("negative", "positive", "all"),
	shade.lwd = 1, shade.col = "white",

	p.mat = NULL, sig.level = 0.05,
	insig = c("pch","p-value","blank", "n"),
	pch = 4, pch.col = "black", pch.cex = 3,
		
	plotCI = c("n","square", "circle", "rect"),
	lowCI.mat = NULL, uppCI.mat = NULL, ...)

Arguments

`corr`	The correlation matrix to visualize, must be square if `order` is not `"original"`. For general matrix, please using `is.corr = FALSE` to convert.
`method`	Character, the visualization method of correlation matrix to be used. Currently, it supports seven methods, named `"circle"` (default), `"square"`, `"ellipse"`, `"number"`, `"pie"`, `"shade"` and `"color"`. See examples for details. The areas of circles or squares show the absolute value of corresponding correlation coefficients. Method `"pie"` and `"shade"` came from Michael Friendly's job (with some adjustment about the shade added on), and `"ellipse"` came from D.J. Murdoch and E.D. Chow's job, see in section References.
`type`	Character, `"full"` (default), `"upper"` or `"lower"`, display full matrix, lower triangular or upper triangular matrix.
`add`	Logical, if `TRUE`, the graph is added to an existing plot, otherwise a new plot is created.
`col`	Vector, the color of glyphs. It is distributed uniformly in `cl.lim`. If NULL, `col` will be `colorRampPalette(col2)(200)`, see example about col2.
`bg`	The background color.
`title`	Character, title of the graph.
`is.corr`	Logical, whether the input matrix is a correlation matrix or not. We can visualize the non-correlation matrix by setting `is.corr = FALSE`.
`diag`	Logical, whether display the correlation coefficients on the principal diagonal.
`outline`	Logical, whether plot outline of circles, square and ellipse.
`mar`	See `par`.
`addgrid.col`	The color of grid. The default value is depends on `method`, if `method` is `color` or `shade`, the default values is `"white"`, otherwise `"grey"`.
`addCoef.col`	Color of coefficients added on the graph. If `NULL` (default), add no coefficients.
`addCoefasPercent`	Logic, whether translate coefficients into percentage style for spacesaving.
`order`	Character, the ordering method of the correlation matrix. `"original"` for original order (default). `"AOE"` for the angular order of the eigenvectors. `"FPC"` for the first principal component order. `"hclust"` for the hierarchical clustering order. `"alphabet"` for alphabetical order. See function `corrMatOrder` for details.
`hclust.method`	Character, the agglomeration method to be used when `order` is `hclust`. This should be one of `"ward"`, `"single"`, `"complete"`, `"average"`, `"mcquitty"`, `"median"` or `"centroid"`.
`addrect`	Integer, the number of rectangles draws on the graph according to the hierarchical cluster, only valid when `order` is `hclust`. If `NULL` (default), then add no rectangles.
`rect.col`	Color for rectangle border(s), only valid when `addrect` is equal or greater than 1.
`rect.lwd`	Numeric, line width for borders for rectangle border(s), only valid when `addrect` is equal or greater than 1.
`tl.pos`	Character or logical, position of text labels. If character, it must be one of `"lt"`, `"ld"`, `"td"`, `"d"` or `"n"`. `"lt"`(default if `type=="full"`) means left and top, `"ld"`(default if `type=="lower"`) means left and diagonal, `"td"`(default if `type=="upper"`) means top and diagonal(near), `"d"` means diagonal, `"n"` means don't add textlabel.
`tl.cex`	Numeric, for the size of text label (variable names).
`tl.col`	The color of text label.
`tl.offset`	Numeric, for text label, see `text`.
`tl.srt`	Numeric, for text label string rotation in degrees, see `text`.
`cl.pos`	Character or logical, position of color labels; If character, it must be one of `"r"` (default if `type=="upper"` or `"full"`), `"b"` (default if `type=="lower"`) or `"n"`, `"n"` means don't draw colorlabel.
`cl.lim`	The limits `(x1, x2)` in the colorlabel.
`cl.length`	Integer, the number of number-text in colorlabel, passed to `colorlegend`. If `NULL`, `cl.length` is `length(col) + 1` when `length(col) <=20`; `cl.length` is 11 when `length(col) > 20`
`cl.cex`	Numeric, cex of number-label in colorlabel, passed to `colorlegend`.
`cl.ratio`	Numeric, to justify the width of colorlabel, 0.1~0.2 is suggested.
`cl.align.text`	Character, `"l"`, `"c"` (default) or `"r"`, for number-label in colorlabel, `"l"` means left, `"c"` means center, and `"r"` means right.
`cl.offset`	Numeric, for number-label in colorlabel, see `text`.
`addshade`	Character for shade style, `"negative"`, `"positive"` or `"all"`, only valid when `method` is `"shade"`. If `"all"`, all correlation coefficients' glyph will be shaded; if `"positive"`, only the positive will be shaded; if `"negative"`, only the negative will be shaded. Note: the angle of shade line is different, 45 degrees for positive and 135 degrees for negative.
`shade.lwd`	Numeric, the line width of shade.
`shade.col`	The color of shade line.
`p.mat`	Matrix of p-value, if `NULL`, arguments `sig.level`, `insig`, `pch`, `pch.col`, `pch.cex` is invalid.
`sig.level`	Significant level, if the p-value in `p-mat` is bigger than `sig.level`, then the corresponding correlation coefficient is regarded as insignificant.
`insig`	Character, specialized insignificant correlation coefficients, `"pch"` (default), `"p-value"`, `"blank"` or `"n"`. If `"blank"`, wipe away the corresponding glyphs; if `"p-value"`, add p-values the corresponding glyphs; if `"pch"`, add characters (see `pch` for details) on corresponding glyphs; if `"n"`, don't take any measures.
`pch`	Add character on the glyphs of insignificant correlation coefficients(only valid when `insig` is `"pch"`). See `par`.
`pch.col`	The color of pch (only valid when `insig` is `"pch"`).
`pch.cex`	The cex of pch (only valid when `insig` is `"pch"`).
`plotCI`	Character, method of ploting confidence interval. If `"n"`, don't plot confidence interval. If "rect", plot rectangles whose upper side means upper bound and lower side means lower bound, respectively, and meanwhile correlation coefficients are also added on the rectangles. If "circle", first plot a circle with the bigger absolute bound, and then plot the smaller. Warning: if the two bounds are the same sign, the smaller circle will be wiped away, thus forming a ring. Method "square" is similar to "circle".
`lowCI.mat`	Matrix of the lower bound of confidence interval.
`uppCI.mat`	Matrix of the upper bound of confidence interval.
`...`	Additional arguments passing to function `text` for drawing text lable.

Details

corrplot function offers flexible ways to visualize correlation matrix, lower and upper bound of confidence interval matrix.

Value

(Invisibly) returns a reordered correlation matrix.

Note

Cairo and cairoDevice packages is strongly recommended to produce high-quality PNG, JPEG, TIFF bitmap files, especially for that method circle, ellipse .

Author(s)

Taiyun Wei

References

Michael Friendly (2002). Corrgrams: Exploratory displays for correlation matrices. The American Statistician, 56, 316–324.

D.J. Murdoch, E.D. Chow (1996). A graphical display of large correlation matrices. The American Statistician, 50, 178–180.

Examples



data(mtcars)
M <- cor(mtcars)
##  different color series
col1 <- colorRampPalette(c("#7F0000","red","#FF7F00","yellow","white", 
			"cyan", "#007FFF", "blue","#00007F"))
col2 <- colorRampPalette(c("#67001F", "#B2182B", "#D6604D", "#F4A582", "#FDDBC7",
			"#FFFFFF", "#D1E5F0", "#92C5DE", "#4393C3", "#2166AC", "#053061"))	
col3 <- colorRampPalette(c("red", "white", "blue"))	
col4 <- colorRampPalette(c("#7F0000","red","#FF7F00","yellow","#7FFF7F", 
			"cyan", "#007FFF", "blue","#00007F"))	
wb <- c("white","black")


par(ask = TRUE)


## different color scale and methods to display corr-matrix
corrplot(M, method="number", col="black", cl.pos="n")
corrplot(M, method="number")
corrplot(M)
corrplot(M, order ="AOE")
corrplot(M, order ="AOE", addCoef.col="grey")

corrplot(M, order="AOE", col=col1(20), cl.length=21,addCoef.col="grey")
corrplot(M, order="AOE", col=col1(10),addCoef.col="grey")

corrplot(M, order="AOE", col=col2(200))
corrplot(M, order="AOE", col=col2(200),addCoef.col="grey")
corrplot(M, order="AOE", col=col2(20), cl.length=21,addCoef.col="grey")
corrplot(M, order="AOE", col=col2(10),addCoef.col="grey")

corrplot(M, order="AOE", col=col3(100))
corrplot(M, order="AOE", col=col3(10))


corrplot(M, method="color", col=col1(20), cl.length=21,order = "AOE", addCoef.col="grey")
corrplot(M, method="square", col=col2(200),order = "AOE")
corrplot(M, method="ellipse", col=col1(200),order = "AOE")
corrplot(M, method="shade", col=col3(20),order = "AOE")
corrplot(M, method="pie", order = "AOE")

## col=wb
corrplot(M, col = wb, order="AOE", outline=TRUE, cl.pos="n")
## like Chinese wiqi, suit for either on screen or white-black print.
corrplot(M, col = wb, bg="gold2",  order="AOE", cl.pos="n")


## mixed methods: It's more efficient if using function "corrplot.mixed"
## circle + ellipse
corrplot(M,order="AOE",type="upper",tl.pos="d")
corrplot(M,add=TRUE, type="lower", method="ell",order="AOE",
	diag=FALSE,tl.pos="n", cl.pos="n")
	
## circle + square
corrplot(M,order="AOE",type="upper",tl.pos="d")
corrplot(M,add=TRUE, type="lower", method="square",order="AOE",
	diag=FALSE,tl.pos="n", cl.pos="n")

## circle + colorful number
corrplot(M,order="AOE",type="upper",tl.pos="d")
corrplot(M,add=TRUE, type="lower", method="number",order="AOE",
	diag=FALSE,tl.pos="n", cl.pos="n")

## circle + black number
corrplot(M,order="AOE",type="upper",tl.pos="tp")
corrplot(M,add=TRUE, type="lower", method="number",order="AOE", col="black",
	diag=FALSE,tl.pos="n", cl.pos="n")

	
## order is hclust and draw rectangles
corrplot(M, order="hclust")
corrplot(M, order="hclust", addrect = 2)
corrplot(M, order="hclust", addrect = 3, rect.col = "red")
corrplot(M, order="hclust", addrect = 4, rect.col = "blue")
corrplot(M, order="hclust", hclust.method="ward", addrect = 4)



## visualize a  matrix in [0, 1]
corrplot(abs(M),order="AOE", cl.lim=c(0,1))
corrplot(abs(M),order="AOE", col=col1(20), cl.lim=c(0,1))
corrplot(abs(M),order="AOE", col=col3(200), cl.lim=c(0,1))


## visualize a  matrix in [-100, 100]
ran <- round(matrix(runif(225, -100,100), 15))
corrplot(ran, is.corr=FALSE)
corrplot(ran, is.corr=FALSE, cl.lim=c(-100, 100))


## text-labels and plot type
corrplot(M, order="AOE", tl.srt=45)
corrplot(M, order="AOE", tl.srt=60)
corrplot(M, order="AOE", tl.pos="d",cl.pos="n")
corrplot(M, order="AOE", diag=FALSE, tl.pos="d")
corrplot(M, order="AOE", type="upper")
corrplot(M, order="AOE", type="upper", diag=FALSE)
corrplot(M, order="AOE", type="lower", cl.pos="b")
corrplot(M, order="AOE", type="lower", cl.pos="b", diag=FALSE)



#### color-legend
corrplot(M, order="AOE", cl.ratio=0.2, cl.align="l")
corrplot(M, order="AOE", cl.ratio=0.2, cl.align="c")
corrplot(M, order="AOE", cl.ratio=0.2, cl.align="r")
corrplot(M, order="AOE", cl.pos="b")
corrplot(M, order="AOE", cl.pos="b", tl.pos="d")
corrplot(M, order="AOE", cl.pos="n")



##the input matrix is not square
corrplot(M[1:8,])
corrplot(M[,1:8])



cor.mtest <- function(mat, conf.level = 0.95){
	mat <- as.matrix(mat)
	n <- ncol(mat)
	p.mat <- lowCI.mat <- uppCI.mat <- matrix(NA, n, n)
	diag(p.mat) <- 0
	diag(lowCI.mat) <- diag(uppCI.mat) <- 1
	for(i in 1:(n-1)){
		for(j in (i+1):n){
			tmp <- cor.test(mat[,i], mat[,j], conf.level = conf.level)
			p.mat[i,j] <- p.mat[j,i] <- tmp$p.value
			lowCI.mat[i,j] <- lowCI.mat[j,i] <- tmp$conf.int[1]
			uppCI.mat[i,j] <- uppCI.mat[j,i] <- tmp$conf.int[2]
		}
	}
	return(list(p.mat, lowCI.mat, uppCI.mat))
}

res1 <- cor.mtest(mtcars,0.95)
res2 <- cor.mtest(mtcars,0.99)


## specialized the insignificant value according to the significant level
corrplot(M, p.mat = res1[[1]], sig.level=0.2)
corrplot(M, p.mat = res1[[1]], sig.level=0.05)
corrplot(M, p.mat = res1[[1]], sig.level=0.01)
corrplot(M, p.mat = res1[[1]], insig = "blank")
corrplot(M, p.mat = res1[[1]], insig = "p-value")
corrplot(M, p.mat = res1[[1]], insig = "p-value", sig.level=-1) ## add all p-values
corrplot(M, p.mat = res1[[1]], order="hclust", insig = "blank", addrect=3)
corrplot(M, p.mat = res1[[1]], order="hclust", insig = "pch", addrect=3)



## plot confidence interval(0.95), "square" method	
corrplot(M,low=res1[[2]], upp=res1[[3]],
	plotC="circle", addg="grey20",cl.pos="n")
corrplot(M, p.mat = res1[[1]],low=res1[[2]], upp=res1[[3]],
	plotC="circle", addg="grey20",cl.pos="n")
corrplot(M, low=res1[[2]], upp=res1[[3]], 
	col=c("white","black"),bg="gold2",order="AOE",
	plotCI="circle",cl.pos="n",pch.col="red")
corrplot(M, p.mat = res1[[1]], low=res1[[2]], upp=res1[[3]], 
	col=c("white","black"),bg="gold2",order="AOE",
	plotCI="circle",cl.pos="n",pch.col="red")
	
## plot confidence interval(0.95), "square" method	
corrplot(M, low=res1[[2]], upp=res1[[3]],
	col=c("white","black"),bg="gold2", order="AOE",
	plotCI="square",addg=NULL,cl.pos="n")
corrplot(M, p.mat = res1[[1]],low=res1[[2]], upp=res1[[3]],
	col=c("white","black"),bg="gold2",order="AOE",pch.col="red", 
	plotC="square", addg=NULL,cl.pos="n")
	
## plot confidence interval(0.95, 0.95, 0.99), "rect" method
corrplot(M, low=res1[[2]], upp=res1[[3]], order="hclust",
	rect.col="navy", plotC="rect",cl.pos="n")
corrplot(M, p.mat = res1[[1]], low=res1[[2]], upp=res1[[3]], order="hclust",
	pch.col="red", sig.level = 0.05, addrect=3, rect.col="navy",
	plotC="rect",cl.pos="n")
corrplot(M, p.mat = res2[[1]], low=res2[[2]], upp=res2[[3]], order="hclust",
	pch.col="red", sig.level = 0.01, addrect=3, rect.col="navy",
	plotC="rect",cl.pos="n")


## an animation of changing confidence interval in different significance level
## begin.animaton
par(ask=FALSE)
for(i in seq(0.1, 0, -0.005)){
	tmp <- cor.mtest(mtcars,1-i)
	corrplot(M, p.mat = tmp[[1]], low=tmp[[2]], upp=tmp[[3]], order="hclust",
		pch.col="red", sig.level = i, plotC="rect", cl.pos="n",
		mar=c(0,0,1,0), 
		title=substitute(alpha == x,list(x=format(i,digits=3,nsmall=3))))
	Sys.sleep(0.15)
}
## end.animaton

[Package corrplot version 0.73 Index]