R/bivariate.R

In DataScienceSalon/xms: Extra-Marital Sex: Attitudes and Behaviors

#==============================================================================#
#                                 bivariate                                    #
#==============================================================================#
#' bivariate
#'
#' \code{bivariate} Performs bivariate analysis of two categorical variables.
#' The analysis includes a chi-square test of association and difference in
#' proportion tests for each level of the categorical explanatory variable,
#' ordered by proportion.
#'
#' @param data Data frame containing response and explanatory variable
#' @param y Character string description for the response variable
#' @param x Character string description of the explanatory variable
#' @param title Character string containing the plot title
#' @param response Character string indicating whether the response is categorical or numeric c("c", "n")
#' @param success Character string for the factor variable considered the success factor
#' @param conf Numeric between 0 and 1 for the designated confidence level
#' @param alpha Numeric between 0 and 1 for the probability of a type I error
#' @param alternative Character string indicating whether the alternative hypothesis is "two.sided", "less", or "greater"
#'
#' @author John James, \email{jjames@@datasciencesalon.org}
#' @family xmar functions
#' @export
bivariate <- function(data,  y, x, title, response = "c", success = "Non-Traditional",
                      conf = 0.95, alpha = 0.05, alternative = "two.sided") {

  if (response == "c") {

    # Create frequency and proportion contingency tables
    freqTbl <- table(data)
    propTbl <- prop.table(table(data), 2)

    # Conduct chi-square test for association
    x2 <- x2Test(freqTbl, y = y, x = x, conf = conf, alpha = alpha)

    # Conduct difference in proportion tests
    dp <- diffTest(data, alternative = alternative,
                   success, conf = conf, alpha = alpha)

    # Plot Data
    analysis <- analyze2DData(x2$htest)
    observed <- plotBars(analysis, y = y, x = x, title = paste("Observed", title))
    x2Plot <- plotX2(x2$result)

    # Format results
    result = list(
      tables = list(
        obsFreq = x2$htest$observed,
        expFreq = x2$htest$expected,
        obsProp = round(prop.table(x2$htest$observed, 2), 3),
        expProp = round(prop.table(x2$htest$expected, 2), 3)
      ),
      analysis = analysis,
      tests = list(
        x2 = x2,
        dp = dp
      ),
      plots = list(
        observed = observed,
        x2Plot = x2Plot
      )
    )
  }
  return(result)
}