R/plot_coefs.R

In m-clark/visibly: Functions and palettes related to visual enhancement

#' Plot coefficients with uncertainty
#'
#' @description This isn't meant to be directly called, but is instead
#'   internally used by the plot_coefficients function.
#'
#' @param model_input Processed model effects
#' @param palette Passed from other methods.
#' @param ref_line Passed from other methods.
#' @param trans Passed from other methods.
#'
#' @return  A ggplot of the coefficients with their corresponding uncertainty
#'   bars.
#'
plot_coefs <- function(
  model_input,
  palette,
  ref_line,
  trans
) {

  if (!requireNamespace("scico", quietly = TRUE)) {
    stop("scico package is needed for this function to work.
         Please install it.",
         call. = FALSE)
  }

  model_input <- model_input %>%
    dplyr::mutate(bold = ifelse(sign(ui_l)*sign(ui_u) == 1, 1, .9),
                  Coefficient = ordered(Coefficient, levels=Coefficient)) # sigh

  if (!is.null(trans))
    model_input <- model_input %>%
      dplyr::mutate_at(vars(value, contains('ui')), trans)

  listcol <- model_input %>%
    split(.$Coefficient) %>%
    purrr::map(function(x) c(seq(x$ui_l, x$value, length.out = 2000),
                             seq(x$value, x$ui_u, length.out = 2000))) %>%
    dplyr::bind_rows() %>%
    tidyr::pivot_longer(
      dplyr::everything(),
      names_to = 'Coefficient',
      values_to = 'ui_value'
    ) %>%
    dplyr::group_by(Coefficient) %>%
    dplyr::mutate(value_sc = abs(scale(ui_value))) %>%
    dplyr::ungroup() %>%
    dplyr::mutate(Coefficient = ordered(Coefficient,
                                        levels = levels(model_input$Coefficient)))
  # sigh again

  pointcol <- scico::scico(n = 1, begin = 1, palette = palette)

  if (pointcol == "#FFFFFF")
    pointcol <- scico::scico(n = 1, begin = 0, end = 0, palette = palette)


  model_input %>%
    ggplot2::ggplot(aes(x = Coefficient, y = value)) +
    ggplot2::geom_hline(yintercept = ref_line, alpha = .1) +
    geom_line(
      aes(group = Coefficient, y = ui_value, color = value_sc),
      data = listcol,
      size = 5,
      show.legend = FALSE
    ) +
    ggplot2::geom_point(size = 3 + .05,
                        # point border
                        color = pointcol,
                        alpha = 1) +
    ggplot2::geom_point(size = 3,
                        color = '#FFFFFF',
                        alpha = 1) +
    ggplot2::geom_point(size = .5,
                        color = pointcol,
                        alpha = 1) +
    scico::scale_color_scico(begin = 1,
                             end = 0,
                             palette = palette) +
    ggplot2::coord_flip() +
    ggplot2::labs(x = '', y = 'Coefficient') +
    theme_clean()


}



plot_coefs_re <- function(model_input,
                          ref_line) {

  model_input <- model_input %>%
    dplyr::mutate(bold = ifelse(sign(ui_l)*sign(ui_u) == 1, 1, .9),
                  Coefficient = factor(Coefficient, levels=Coefficient))  # sigh

  model_input %>%
    ggplot2::ggplot(aes(x = Coefficient, y=value)) +
    ggplot2::geom_hline(yintercept = ref_line, alpha = .25, color='#ff5500') +
    geom_linerange(aes(group=Coefficient, ymin=ui_l, ymax=ui_u, alpha=bold),
              color = 'gray50',
              size=1,
              show.legend = FALSE) +
    ggplot2::geom_point(size=3 + .05,     # point border
                        color = 'lightblue',
                        alpha=1) +
    ggplot2::geom_point(size=3,
                        color = '#FFFFFF',
                        alpha=1) +
    ggplot2::geom_point(size=.5,
                        color = 'lightblue',
                        alpha=1) +
    ggplot2::labs(x='', y='Coefficient') +
    theme_clean()


}