R/phackr.R

In willdebras/phackr: Produces excel sheets showing multivariate significance and directionality

#' phackr
#'
#'The phackr function takes a dataset, a vector of dependent variables, and a vector of demographic variables and creates a tibble of multivariate significance from ordinal logit models.
#'
#' @param data A survey data object. This can be created with the phackr_setup function
#' @param dvs A vector of dependent variables to run the same model across.
#' @param demos A vector of demographic variables and covariates to include in the ologit model.
#' @param logit A parameter to run logistic regressions. Defaults to ordinal logistic regression.
#'
#' @return Returns a tibble that can quickly be written to an excel sheet
#' @export
#' @importFrom survey svyolr svyglm
#' @importFrom broom tidy
#' @import tibble
#' @import magrittr
#'
#' @examples
#' sheet1 <- phackr(data = survey_data, dvs = c("q2", "q3", "q4"), demos = c("race", "partyid", "gender"))
#'
phackr <- function(data, dvs, demos, logit = FALSE) {

  ###create empty lists to fill
  dv_models <- list()

  dv <- as.list(dvs)

  if (logit) {

    for (i in dv) {

      dv_models[[i]] <- tidy(svyglm(paste(i[[1]], paste("~ ", paste(demos, collapse = " + "))), family = quasibinomial, data))

    }

    sheet <- tibble(rows = demos)

    for (i in 1:length(dv_models)) {

      sheet[(i+1)] <- add_column(extract_elements_logit_single(dv_models, i))

    }

  }

  else {

    for (i in dv) {

      dv_models[[i]] <- tidy(svyolr(paste(i[[1]], paste("~ ", paste(demos, collapse = " + "))), data))

    }
    sheet <- tibble(rows = demos)

    for (i in 1:length(dv_models)) {

      sheet[(i+1)] <- add_column(extract_elements_ologit_single(dv_models, i))

    }

  }

  sheet <- sheet %>%
    remove_rownames %>%
    column_to_rownames(var="rows") %>%
    `colnames<-`(unlist(dv, use.names = FALSE))

  return(sheet)

}