R/pairsContVars.R

In CDCgov/DemographySpawnR: Simulate A New Dataset Based on Distribution of a Given Dataset

#' Pairs of all continuous variables that may be associated
#'
#' @param inputData Dataset that you want investigate
#'
#' @return list of all pairs of continuous variables that may be correlated
#'
#' @export
#'
#' @examples
#' pairsContVars()
#'
#'
pairsContVars = function (inputData) {


  inputData = data.frame(inputData)
  corPvals = matrix(nrow = ncol(inputData)**2, ncol = 4)

  nrowsPval = 0

  for (i in 1:ncol(inputData)) {

    for (j in 1:ncol(inputData)) {

      if (i == j) next

      if ((is.numeric(inputData[,i]) & length(unique(inputData[,i])) > 5) &
          (is.numeric(inputData[,j]) & length(unique(inputData[,j])) > 5)) {

        correlation = cor.test(inputData[,i], inputData[,j])

        nrowsPval = nrowsPval + 1;


        corPvals[nrowsPval, 1] = paste0(names(inputData)[i], ", ", names(inputData)[j])
        corPvals[nrowsPval, 2] = round(correlation$estimate, 3)
        corPvals[nrowsPval, 3] = paste0("(",round(correlation$conf.int[1],3), ", ",
                                        round(correlation$conf.int[2],3), ")")

        corPvals[nrowsPval, 4] = round(correlation$p.value, 3)


        corPvals2 = as.data.frame(corPvals[!duplicated(corPvals[,c(2,3,4)]),])

        names(corPvals2) = c( "contVarsTested", "corrEstimate", "CI95", "pValue")

        corPvals2 = subset(corPvals2, is.na(corPvals2$corrEstimate) == F)

        pairs = as.data.frame(corPvals2$contVarsTested)
        names(pairs) = "catVarsTested"

      } # close the if statement looking for numeric/continuous data
    } # close j for loop
  } # close i for loop
  return(pairs)
} # close function