R/densityPlot.R

In cardiomoon/processR: Implementation of the 'PROCESS' Macro

#' Get bootstrapped values
#' @param semfit An object of class lavaan
#' @param what Character. What needs to be inspected/extracted?
#' @param ... Further argument to be passed to lavaan::lavTech()
#' @importFrom lavaan lavTech
#' @importFrom utils packageVersion
#' @export
#' @examples
#' require(lavaan)
#' labels=list(X="cond",M="pmi",Y="reaction")
#' model=tripleEquation(labels=labels)
#' \donttest{
#' set.seed(1234)
#' semfit=sem(model,data=pmi,se="boot",bootstrap=100)
#' getBootData(semfit)
#' }
getBootData=function(semfit,what="coef.boot",...){
    if(packageVersion("lavaan")<"0.6.5.1453") {
        cat("To use this function, please install the latest 'lavaan' package using the following R code.\n")
        cat("install.packages('lavaan',repos='http://www.da.ugent.be',type='source')\n")
    }
    as.data.frame(lavTech(semfit, what=what, add.labels = TRUE,...))

}


#' Draw Smoothed Kernel density plot
#' @param x A numeric vector
#' @param sig significant level. Default value is 0.05
#' @param digits Integer indicating the number of decimal places
#' @param xlab character. x axis label
#' @param ylab character. y axis label
#' @importFrom ggplot2 ggplot geom_histogram annotate labs geom_line
#' @importFrom predict3d theme_bw2
#' @export
#' @examples
#' require(lavaan)
#' labels=list(X="cond",M="pmi",Y="reaction")
#' model=tripleEquation(labels=labels)
#' \donttest{
#' set.seed(1234)
#' semfit=sem(model,data=pmi,se="boot",bootstrap=100)
#' bootData=getBootData(semfit)
#' bootData$indirect=bootData$a*bootData$b
#' densityPlot(bootData$indirect)
#' }
densityPlot=function(x,sig=0.05,digits=3,xlab="Indirect effect(ab)",ylab=NULL){
    if(is.null(ylab)){
        ylab=paste0("Smoothed Kernel density estimates \nin ",length(x)," bootstrap samples")
    }
    df=as.data.frame(x=x)
    xintercept=quantile(x,probs=c(sig/2,1-sig/2),type=6,na.rm=TRUE)
    labels=sprintf(paste0("%0.",digits,"f"),xintercept)

    p<-ggplot(data=df,aes_string(x="x",y="..density.."))+
        geom_histogram(color="grey60",fill="cornsilk")+
        geom_line(stat="density",color="red")

    res=getAspectRatio(p)

    p<-p+ geom_vline(xintercept=xintercept[1],lty=2)+
        geom_vline(xintercept=xintercept[2],lty=2)+
        annotate("text",x=xintercept[1],y=res$ymax,label=labels[1],hjust=1.1)+
        annotate("text",x=xintercept[2],y=res$ymax,label=labels[2],hjust=-0.1)+
        labs(x=xlab,y=ylab)+
        theme_bw2()
    p
}

#' Draw quantile-quantile plot
#' @param x A numeric vector
#' @param linecolor character line color
#' @param xlab character label for x axis
#' @param ylab character label for y axis
#' @param title character label for plot title
#' @param ... Further arguments to be passed to geom_qq()
#' @export
#' @importFrom ggplot2 geom_qq geom_qq_line labs element_text
#' @examples
#' qqPlot(rnorm(200))
#' qqPlot(rt(200, df = 5))
qqPlot=function(x,linecolor="red",xlab=NULL,ylab=NULL,title=NULL,...){
    if(is.null(xlab)) xlab="Theoretical Quantiles"
    if(is.null(ylab)) ylab="Sample Quantiles"
    if(is.null(title)) title="Normal Q-Q Plot"
    ggplot(data=as.data.frame(x),aes(sample=x))+geom_qq(...)+
        geom_qq_line(col=linecolor)+
        theme_bw2()+
        labs(title=title,x=xlab,y=ylab)+
        theme(plot.title=element_text(hjust=0.5))
}