R/draw_multivari_plot.R

In gaboraszabo/sherlock: Graphical Displays for Structured Problem Solving and Diagnosis

#' Draw Multivari Plot
#'
#' @description
#' Draws a multivari small multiples plot
#'
#' @param data Input dataset to be plotted (required)
#' @param y_var Response variable, Y (required)
#' @param grouping_var_1 Select column for lowest level grouping variable (required)
#' @param grouping_var_2 Select column for second level grouping variable (required)
#' @param grouping_var_3 Select column for third level grouping variable (optional)
#' @param grouping_var_4 Select column for fourth level grouping variable (optional)
#' @param unnest_grouping_var_2 Unnest second level grouping variable. Useful when characterizing shape/geometry (optional)
#' @param data_point_label Select column to label data points (optional)
#' @param plot_means Logical. if FALSE, default, means for mid-level factor are not plotted (optional)
#' @param point_size Set point size. By default, it is set to 2.5  (optional)
#' @param line_size Set line size. By default, it is set to 0.7  (optional)
#' @param alpha Set transparency. By default, it is set to 0.6  (optional)
#' @param x_axis_text_size Set X axis text size. By default, it is set to 11  (optional)
#' @param panel_text_size Set panel text size. By default, it is set to 14 (optional)
#'
#' @return A 'ggplot' object
#'
#' @examples
#' library(dplyr)
#' library(ggh4x)
#'
#' polar_small_multiples_data %>%
#'   filter(ID_Measurement_Angle %in% c(0, 45, 90, 135)) %>%
#'   normalize_observations(y_var = ID,
#'                          grouping_var = Tip_Bottom,
#'                          ref_values = c(0.2075, 0.2225)) %>%
#'   draw_multivari_plot(y_var            = ID_normalized,
#'                      grouping_var_1    = ID_Measurement_Angle,
#'                      grouping_var_2    = Mold_Cavity_Number,
#'                      grouping_var_3    = Tip_Bottom,
#'                      x_axis_text = 6) +
#'   draw_horizontal_reference_line(reference_line = 0)
#'
#' @export


draw_multivari_plot <- function(data, y_var, grouping_var_1, grouping_var_2, grouping_var_3, grouping_var_4, unnest_grouping_var_2 = FALSE,
                                data_point_label = NULL, plot_means = FALSE, x_axis_text_size = 11, panel_text_size = 14, point_size = 2.5,
                                line_size = 0.7, alpha = 0.6) {

  # 1. Tidy Eval ----
  response_expr   <- rlang::enquo(y_var)
  factor_1_expr   <- rlang::enquo(grouping_var_1)
  factor_2_expr   <- rlang::enquo(grouping_var_2)
  factor_3_expr   <- rlang::enquo(grouping_var_3)
  factor_4_expr   <- rlang::enquo(grouping_var_4)
  data_point_label_expr <- rlang::enquo(data_point_label)

  # 2. Theme element ----
  theme_element <- ggplot2::theme_light() +
    ggplot2::theme(
      panel.grid.major  = ggplot2::element_blank(),
      panel.grid.minor  = ggplot2::element_blank(),
      panel.spacing     = ggplot2::unit(0, "lines"),
      panel.border      = ggh4x::element_part_rect(color = "grey95", linewidth = 0.2),
      strip.background  = ggplot2::element_rect(fill = "white", color = "grey70", linewidth = 0.5),
      strip.text        = ggplot2::element_text(size = panel_text_size, color = "grey50"),
      plot.title        = ggplot2::element_text(size = 20, color = "grey50"),
      plot.subtitle     = ggplot2::element_text(size = 14, color = "grey70"),
      axis.title.x      = ggplot2::element_text(size = 13, color = "grey70"),
      axis.title.y      = ggplot2::element_text(size = 13, color = "grey70"),
      axis.text.x       = ggplot2::element_text(size = x_axis_text_size, color = "grey70"),
      axis.text.y       = ggplot2::element_text(size = 12, color = "grey70"),
      axis.line.x.bottom        = ggplot2::element_line(color = "grey70", size = 0.5),
      axis.line.y.left  = ggplot2::element_line(color = "grey70", size = 0.5),
      axis.line.y.right = ggplot2::element_line(color = "grey70", size = 0.5),
      axis.text.y.right = ggplot2::element_blank(),
      legend.position   = "none",
      plot.caption      = ggplot2::element_text(color = "grey60"),
      axis.ticks        = ggplot2::element_blank()
    )

  # 3. Warning messages ----
  if (missing(grouping_var_2) & missing(grouping_var_3)) message("You need two specify at least two grouping variables (grouping_var_1 and grouping_var_2)!")

  if (!((data %>% dplyr::pull(!!response_expr)) %>% is.numeric())) {
    message("Response variable is not numeric. Are you sure this is what you meant to do?")
  }


  # 4. Conditionals ----
  # 4.1 if only 2 factors ----
  if (missing(grouping_var_3)) {

    # Mean tbl for plotting averages ----
    mean_tbl <- data %>%
      dplyr::mutate(!!(factor_2_expr) := forcats::as_factor(!!factor_2_expr)) %>%
      dplyr::mutate(!!(factor_1_expr) := forcats::as_factor(!!factor_1_expr)) %>%
      dplyr::group_by(!!factor_2_expr, !!factor_1_expr) %>%
      dplyr::summarize(mean = mean(!!response_expr)) %>%
      dplyr::ungroup()


    # Multi-vari tbl ----
    multi_vari_tbl <- data %>%
      dplyr::mutate(!!(factor_2_expr) := forcats::as_factor(!!factor_2_expr)) %>%
      dplyr::mutate(!!(factor_1_expr) := forcats::as_factor(!!factor_1_expr))


    # Plotting function ----
    multi_vari_chart <- multi_vari_tbl %>%

      ggplot2::ggplot(ggplot2::aes(!!factor_1_expr, !!response_expr, color = !!factor_2_expr)) +
      # ggh4x::facet_nested(rows = ggplot2::vars(), cols = ggplot2::vars(!!factor_2_expr)) +
      ggplot2::guides(y.sec = "axis") +
      theme_element +
      ggplot2::labs(
        title = "Multivari Plot",
        subtitle = stringr::str_glue("{as_label(factor_1_expr)} by {as_label(factor_2_expr)}"),
        x = stringr::str_glue("{as_label(factor_1_expr)}"),
        y = stringr::str_glue("{as_label(response_expr)}")
      )

    if (plot_means) {
      multi_vari_chart <- multi_vari_chart +
        ggplot2::geom_point(color = "grey30", size = 1.5, alpha = 0.4) +
        ggplot2::geom_point(data = mean_tbl, ggplot2::aes(!!factor_1_expr, mean), color = "#40506e", size = 3, alpha = 0.7) +
        ggplot2::geom_line(data = mean_tbl, ggplot2::aes(!!factor_1_expr, mean, group = 1), color = "#40506e", alpha = 0.7) +
        ggh4x::facet_nested(rows = ggplot2::vars(), cols = ggplot2::vars(!!factor_2_expr)) +
        ggplot2::labs(caption  = stringr::str_glue("Blue data points represent averages for grouping variable {as_label(factor_1_expr)}"))
    } else {
      multi_vari_chart <- multi_vari_chart +
        ggplot2::geom_point(size = point_size, alpha = alpha) +
        ggplot2::geom_line(ggplot2::aes(group = !!factor_2_expr), linewidth = line_size, alpha = alpha)
      #ggh4x::facet_nested(rows = ggplot2::vars(), cols = ggplot2::vars(!!factor_2_expr))

      if (unnest_grouping_var_2) {
        multi_vari_chart <- multi_vari_chart
      } else {
        multi_vari_chart <- multi_vari_chart +
          ggh4x::facet_nested(rows = ggplot2::vars(), cols = ggplot2::vars(!!factor_2_expr))
      }


      if (!missing(data_point_label)) {
        multi_vari_chart <- multi_vari_chart +
          ggplot2::geom_label((ggplot2::aes(label = !!data_point_label_expr)), size = 3, color = "grey50") +
          ggplot2::labs(
            title = "Multivari Plot",
            subtitle = stringr::str_glue("{as_label(factor_1_expr)} by {as_label(factor_2_expr)}, data point label: {as_label(data_point_label_expr)}"),
            x = stringr::str_glue("{as_label(factor_1_expr)}"),
            y = stringr::str_glue("{as_label(response_expr)}")
          )
      }

    }
  }

  # 4.2 if 3 factors ----
  if (!missing(grouping_var_3) & missing(grouping_var_4)) {

    # Mean tbl for plotting averages ----
    mean_tbl <- data %>%
      dplyr::mutate(!!(factor_3_expr) := forcats::as_factor(!!factor_3_expr)) %>%
      dplyr::mutate(!!(factor_2_expr) := forcats::as_factor(!!factor_2_expr)) %>%
      dplyr::group_by(!!factor_3_expr, !!factor_2_expr) %>%
      dplyr::summarize(mean = mean(!!response_expr)) %>%
      dplyr::ungroup()

    # Multi-vari tbl ----
    multi_vari_tbl <- data %>%
      dplyr::mutate(!!(factor_3_expr) := forcats::as_factor(!!factor_3_expr)) %>%
      dplyr::mutate(!!(factor_2_expr) := forcats::as_factor(!!factor_2_expr)) %>%
      dplyr::mutate(!!(factor_1_expr) := forcats::as_factor(!!factor_1_expr))

    # Plotting function ----
    if (plot_means) {

      multi_vari_chart <- multi_vari_tbl %>%
        ggplot2::ggplot(ggplot2::aes(!!factor_2_expr, !!response_expr)) +
        ggplot2::geom_point(color = "grey30", size = 1.5, alpha = 0.4) +
        ggplot2::geom_point(data = mean_tbl, ggplot2::aes(!!factor_2_expr, mean), color = "#40506e", size = 3, alpha = 0.7) +
        ggplot2::geom_line(data = mean_tbl, ggplot2::aes(!!factor_2_expr, mean, group = 1), color = "#40506e", alpha = 0.7) +

        ggh4x::facet_nested(rows = ggplot2::vars(), cols = ggplot2::vars(!!factor_3_expr)) +
        ggplot2::guides(y.sec = "axis") +

        theme_element +
        ggplot2::labs(
          title    = "Multivari Plot",
          subtitle = stringr::str_glue("{as_label(factor_1_expr)} by {as_label(factor_2_expr)} by {as_label(factor_3_expr)}"),
          x        = stringr::str_glue("{as_label(factor_2_expr)}"),
          y        = stringr::str_glue("{as_label(response_expr)}"),
          caption  = stringr::str_glue("Blue data points represent averages for grouping variable {as_label(factor_2_expr) }
                                       Grey data points represent individual values for grouping variable {as_label(factor_1_expr)}")
        )

    } else {

      multi_vari_chart <- multi_vari_tbl %>%

        ggplot2::ggplot(ggplot2::aes(!!factor_1_expr, !!response_expr, color = !!factor_2_expr, group = !!factor_2_expr)) +
        ggplot2::geom_point(size = point_size, alpha = alpha) +
        ggplot2::geom_line(linewidth = line_size, alpha = alpha)


      if (unnest_grouping_var_2) {
        multi_vari_chart <- multi_vari_chart +
          ggh4x::facet_nested(rows = ggplot2::vars(), cols = ggplot2::vars(!!factor_3_expr))

      } else {
        multi_vari_chart <- multi_vari_chart +
          ggh4x::facet_nested(rows = ggplot2::vars(), cols = ggplot2::vars(!!factor_3_expr,
                                                                           !!factor_2_expr))
      }

      multi_vari_chart <- multi_vari_chart +

        ggplot2::guides(y.sec = "axis") +

        theme_element +
        ggplot2::labs(
          title    = "Multivari Plot",
          subtitle = stringr::str_glue("{as_label(factor_1_expr)} by {as_label(factor_2_expr)} by {as_label(factor_3_expr)}"),
          x        = stringr::str_glue("{as_label(factor_1_expr)}"),
          y        = stringr::str_glue("{as_label(response_expr)}")
        )


      if (!missing(data_point_label)) {
        multi_vari_chart <- multi_vari_chart +
          ggplot2::geom_label((ggplot2::aes(label = !!data_point_label_expr)), size = 3, color = "grey50") +
          ggplot2::labs(
            title    = "Multivari Plot",
            subtitle = stringr::str_glue("{as_label(factor_1_expr)} by {as_label(factor_2_expr)} by {as_label(factor_3_expr)}, data point label: {as_label(data_point_label_expr)}"),
            x        = stringr::str_glue("{as_label(factor_1_expr)}"),
            y        = stringr::str_glue("{as_label(response_expr)}")
          )
      }

    }

  }

  # 4.2 if 4 factors ----
  if (!missing(grouping_var_3) & !missing(grouping_var_4)) {

    # Mean tbl for plotting averages ----
    mean_tbl <- data %>%
      dplyr::mutate(!!(factor_4_expr) := forcats::as_factor(!!factor_4_expr)) %>%
      dplyr::mutate(!!(factor_3_expr) := forcats::as_factor(!!factor_3_expr)) %>%
      dplyr::mutate(!!(factor_2_expr) := forcats::as_factor(!!factor_2_expr)) %>%
      dplyr::group_by(!!factor_4_expr, !!factor_3_expr, !!factor_2_expr) %>%
      dplyr::summarize(mean = mean(!!response_expr)) %>%
      dplyr::ungroup()

    # Multi-vari tbl ----
    multi_vari_tbl <- data %>%
      dplyr::mutate(!!(factor_4_expr) := forcats::as_factor(!!factor_4_expr)) %>%
      dplyr::mutate(!!(factor_3_expr) := forcats::as_factor(!!factor_3_expr)) %>%
      dplyr::mutate(!!(factor_2_expr) := forcats::as_factor(!!factor_2_expr)) %>%
      dplyr::mutate(!!(factor_1_expr) := forcats::as_factor(!!factor_1_expr))

    # Plotting function ----
    if (plot_means) {

      multi_vari_chart <- multi_vari_tbl %>%
        ggplot2::ggplot(ggplot2::aes(!!factor_2_expr, !!response_expr)) +
        ggplot2::geom_point(color = "grey30", size = 1.5, alpha = 0.4) +
        ggplot2::geom_point(data = mean_tbl, ggplot2::aes(!!factor_2_expr, mean), color = "#40506e", size = 3, alpha = 0.7) +
        ggplot2::geom_line(data = mean_tbl, ggplot2::aes(!!factor_2_expr, mean, group = 1), color = "#40506e", alpha = 0.7) +

        ggh4x::facet_nested(rows = ggplot2::vars(), cols = ggplot2::vars(!!factor_4_expr, !!factor_3_expr)) +
        ggplot2::guides(y.sec = "axis") +

        theme_element +
        ggplot2::labs(
          title    = "Multivari Plot",
          subtitle = stringr::str_glue("{as_label(factor_1_expr)} by {as_label(factor_2_expr)} by {as_label(factor_3_expr)} by {as_label(factor_4_expr)}"),
          x        = stringr::str_glue("{as_label(factor_2_expr)}"),
          y        = stringr::str_glue("{as_label(response_expr)}"),
          caption  = stringr::str_glue("Blue data points represent averages for grouping variable {as_label(factor_2_expr) }
                                       Grey data points represent individual values for grouping variable {as_label(factor_1_expr)}")
        )

    } else {

      multi_vari_chart <- multi_vari_tbl %>%

        ggplot2::ggplot(ggplot2::aes(!!factor_1_expr, !!response_expr, color = !!factor_2_expr, group = !!factor_2_expr)) +
        ggplot2::geom_point(size = point_size, alpha = alpha) +
        ggplot2::geom_line(linewidth = line_size, alpha = alpha)

      if (unnest_grouping_var_2) {
        multi_vari_chart <- multi_vari_chart +
          ggh4x::facet_nested(rows = ggplot2::vars(), cols = ggplot2::vars(!!factor_4_expr,
                                                                           !!factor_3_expr,))
      } else {
        multi_vari_chart <- multi_vari_chart +
          ggh4x::facet_nested(rows = ggplot2::vars(), cols = ggplot2::vars(!!factor_4_expr,
                                                                           !!factor_3_expr,
                                                                           !!factor_2_expr))
      }

      multi_vari_chart <- multi_vari_chart +

        ggplot2::guides(y.sec = "axis") +

        theme_element +
        ggplot2::labs(
          title    = "Multivari Plot",
          subtitle = stringr::str_glue("{as_label(factor_1_expr)} by {as_label(factor_2_expr)} by {as_label(factor_3_expr)} by {as_label(factor_4_expr)}"),
          x        = stringr::str_glue("{as_label(factor_1_expr)}"),
          y        = stringr::str_glue("{as_label(response_expr)}")
        )


      if (!missing(data_point_label)) {
        multi_vari_chart <- multi_vari_chart +
          ggplot2::geom_label((ggplot2::aes(label = !!data_point_label_expr)), size = 3, color = "grey50") +
          ggplot2::labs(
            title    = "Multivari Plot",
            subtitle = stringr::str_glue("{as_label(factor_1_expr)} by {as_label(factor_2_expr)} by {as_label(factor_3_expr)} by {as_label(factor_4_expr)}, data point label: {as_label(data_point_label_expr)}"),
            x        = stringr::str_glue("{as_label(factor_1_expr)}"),
            y        = stringr::str_glue("{as_label(response_expr)}")
          )
      }

    }

  }


  if (unnest_grouping_var_2) {
    multi_vari_chart <- multi_vari_chart +
      ggplot2::theme(legend.position = "right",
            legend.title    = ggplot2::element_text(size = 14, color = "grey70"),
            legend.text     = ggplot2::element_text(size = 14, color = "grey70"))
  }


  # 5. Scale color ----
  multi_vari_chart <- multi_vari_chart +
    sherlock::scale_color_sherlock()


  return(multi_vari_chart)

}