R/see_num_vars.R

In sarazong/SZRtools: Sara Zong's Customized Data Analysis Tools

#' @title Histogram for Numeric Variable
#' @description Create a histogram to display the distribution of a numeric variable.
#'
#' @param data dataset that contains the variables
#' @param num_var name of the numeric variable as a string
#' @param bins number of bins, default to 30
#' @param title title of the plot
#'
#' @return none
#' @examples
#' make_histogram(ggplot2::diamonds, "price")
#' @export

make_histogram <- function(data, num_var, bins = 30, title = "") {
  xlab <- stringr::str_to_title(stringr::str_replace(num_var, "[:punct:]", " "))

  ggplot2::ggplot(data, ggplot2::aes(x = get(num_var))) +
    ggplot2::geom_histogram(ggplot2::aes(y = ..density..),
                            bins = bins,
                            color = "black",
                            fill = "grey") +
    ggplot2::geom_density(fill = "red", alpha = 0.2) +
    ggplot2::theme_classic() +
    ggplot2::labs(x = xlab, title = title)
}



#' @title Boxplot for Numeric Grouped by Categorical
#' @description Create a boxplot to display the distribution of a numeric variable
#' grouped by a categorical variable. Good for categorical variable that has ten
#' or less categories. When the maximum observations for a category is < 200, each
#' observation will be plotted and outliers will be hidden, vice versa for
#' observations >= 200. Median for each category is labeled in the box.
#'
#' @param data dataset that contains the variables
#' @param num_var name of the numeric variable as a string
#' @param cat_var name of the categorical variable as a string
#' @param plot_title title for the plot as a string
#' @param vjust position of the median labels
#'
#' @return none
#' @examples
#' make_boxplot1(ggplot2::diamonds, "price", "cut", "Price of Diamonds by Cuts")
#' @export

make_boxplot1 <- function(data, num_var, cat_var, plot_title, vjust = 1.5) {
  num_cat <- szrtools::count_cats(data, c(cat_var))

  if (num_cat > 10) {
    stop("\U0001F611 Too many categories, need a customized boxplot!")
  } else {
    font_size <- switch(as.character(num_cat),
                        "2" = 4.3, "3" = 4.0, "4" = 3.7, "5" = 3.4,
                        "6" = 3.1, "7" = 2.8, "8" = 2.5, "9" = 2.1, "10" = 2)
    # "[:punct:]" are punctuation characters
    ylab <- stringr::str_to_title(stringr::str_replace(num_var, "[:punct:]", " "))
    xlab <- stringr::str_to_title(stringr::str_replace(cat_var, "[:punct:]", " "))

    obs <- szrtools::count_obs(data, c(cat_var), max)

    p <- ggplot2::ggplot(data, ggplot2::aes(x = get(cat_var),
                                            y = get(num_var),
                                            color = get(cat_var)))
    if (obs < 200) {
      p <- p + ggplot2::geom_boxplot(outlier.shape = NA) +
        ggplot2::geom_jitter(alpha = 0.3)
    } else {
      p <- p + ggplot2::geom_boxplot()
    }

    p + ggplot2::stat_summary(ggplot2::aes(label = ggplot2::after_stat(y)), fun = "median",
                              geom = "text", size = font_size,
                              color = "black", vjust = vjust) +
      ggplot2::theme_classic() +
      ggplot2::theme(legend.position = "none") +
      ggplot2::labs(x = xlab, y = ylab, title = plot_title)
  }
}



#' @title Boxplot for Numeric Grouped by Two Categoricals
#' @description Create a boxplot to compare the distribution of a numeric variable
#' grouped by a categorical variable among the different levels of another categorical
#' variable. Good for categorical variables that has ten or less categories. When
#' the maximum observations for a category is < 200, each observation will be plotted
#' and outliers will be hidden, vice versa for observations >= 200.
#'
#' @param data dataset that contains the variables
#' @param num_var name of the numeric variable as a string
#' @param cat_var name of the categorical variable as a string
#' @param grp_var name of the categorical variable as a string
#' @param label_fmt boolean to indicate whether plot labels should be formatted
#'
#' @return none
#' @examples
#' make_boxplot2(ggplot2::diamonds, "price", "cut", "color")
#' @export

make_boxplot2 <- function(data, num_var, cat_var, grp_var, label_fmt = TRUE) {
  num_cat1 <- szrtools::count_cats(data, c(cat_var))

  if (num_cat1 > 7) {
    stop("\U0001F611 Too many categories, need a customized boxplot!")
  } else {
    num_cat2 <- szrtools::count_cats(data, c(grp_var))
    total <- num_cat1 * num_cat2

    per_row <- switch(as.character(num_cat1),
                      "2" = 6, "3" = 5, "4" = 4, "5" = 3, "6" = 2, "7" = 2)
    num_rows <- ceiling(num_cat2 / per_row)

    obs <- szrtools::count_obs(data, c(cat_var, grp_var), max)

    labs <- label_plot(plot_type = "Boxplot", num_var, cat_var, grp_var,
                       label_fmt = label_fmt)

    p <- ggplot2::ggplot(data, ggplot2::aes(x = get(cat_var),
                                            y = get(num_var),
                                            color = get(cat_var)))

    if (total <= 16) {
      if (obs < 200) {
        p <- p + ggplot2::geom_boxplot(outlier.shape = NA) +
          ggplot2::geom_jitter(alpha = 0.3) +
          ggplot2::facet_grid(~ get(grp_var))
      } else {
        p <- p + ggplot2::geom_boxplot() +
          ggplot2::facet_grid(~ get(grp_var))
      }
    } else {
      if (obs < 200) {
        p <- p + ggplot2::geom_boxplot(outlier.shape = NA) +
          ggplot2::geom_jitter(alpha = 0.3) +
          ggplot2::facet_wrap(~ get(grp_var), nrow = num_rows)
      } else {
        p <- p + ggplot2::geom_boxplot() +
          ggplot2::facet_wrap(~ get(grp_var), nrow = num_rows)
      }
    }

    p + ggplot2::theme_classic() +
      ggplot2::theme(axis.text.x = ggplot2::element_text(angle = 45, hjust = 1)) +
      ggplot2::labs(x = labs[[1]][[3]], y = labs[[1]][[1]],
                    title = labs[[2]][[1]], color = labs[[1]][[2]])
      # ggplot2::guides(color = ggplot2::guide_legend(title = legend))
      # ggplot2::annotate("segment", x = -Inf, xend = Inf, y = -Inf, yend = -Inf, size = 1)
  }
}



#' @title Scatter Plot for Numeric Vars
#' @description Create a scatter plot to examine relationship between two numeric
#' variables.
#'
#' @param data dataset that contains the variables
#' @param num_var1 name of the numeric variable as a string
#' @param num_var2 name of the numeric variable as a string
#' @param trend string represents smoothing method to use
#' @param label_fmt Boolean to indicate whether plot labels should be formatted
#'
#' @return none
#' @examples
#' make_scatter_plot1(ggplot2::diamonds, "carat", "price")
#' @export

make_scatter_plot1 <- function(data, num_var1, num_var2,
                               trend = "lm", label_fmt = TRUE) {
  obs <- nrow(data)
  alpha <- dplyr::case_when(obs >= 10000 ~ 0.1, obs >= 5000 ~ 0.25, obs >= 2500 ~ 0.3,
                            obs >= 1000 ~ 0.4, obs >= 200 ~ 0.5, TRUE ~ 0.7)

  labs <- label_plot(plot_type = "Scatter Plot", num_var1, num_var2,
                     label_fmt = label_fmt)

  ggplot2::ggplot(data, ggplot2::aes(x = get(num_var1), y = get(num_var2))) +
    ggplot2::geom_point(alpha = alpha) +
    ggpubr::stat_cor(p.accuracy = 0.001, r.accuracy = 0.01,
                     label.x.npc = 0.7, label.y.npc = 0) +
    ggplot2::geom_smooth(method = trend, se = FALSE, color = "blue") +
    ggplot2::theme_bw() +
    ggplot2::labs(x = labs[[1]][[1]], y = labs[[1]][[2]],
                  title = labs[[2]][[1]])
}



#' @title Scatter Plot for Numeric Grouped by Categorical
#' @description Create a scatter plot to examine relationship between two numeric
#' variables within the categories of a categorical variable.
#'
#' @param data dataset that contains the variables
#' @param num_var1 name of the numeric variable as a string
#' @param num_var2 name of the numeric variable as a string
#' @param grp_var name of the categorical variable as a string
#' @param trend string represents smoothing method to use
#' @param split Boolean to indicate whether categories of the grouping variable
#' should be split into individual plot
#' @param label_fmt Boolean to indicate whether plot labels should be formatted
#'
#' @return none
#' @examples
#' make_scatter_plot2(ggplot2::diamonds, "carat", "price", "cut")
#' @export

make_scatter_plot2 <- function(data, num_var1, num_var2, grp_var, trend = "lm",
                               split = FALSE, label_fmt = TRUE) {

  obs <- nrow(data)
  alpha <- dplyr::case_when(obs >= 10000 ~ 0.25, obs >= 5000 ~ 0.3, obs >= 2500 ~ 0.4,
                            obs >= 1000 ~ 0.5, obs >= 200 ~ 0.6, TRUE ~ 0.7)

  labs <- label_plot(plot_type = "Scatter Plot", num_var1, num_var2, grp_var,
                     label_fmt = label_fmt)

  p <- ggplot2::ggplot(data, ggplot2::aes(x = get(num_var1),
                                     y = get(num_var2),
                                     color = get(grp_var))) +
    ggplot2::geom_point(alpha = alpha)

  if (split) {
    p <-  p + ggpubr::stat_cor(aes(label = ..r.label..), r.accuracy = 0.01,
                               color = "black", size = 2) +
      ggplot2::geom_smooth(method = trend, se = FALSE, color = "black") +
      ggplot2::facet_wrap(~ get(grp_var), ncol = 2, scales = "free_y") +
      ggplot2::guides(color = ggplot2::guide_legend(override.aes = list(alpha = 1)))
  } else {
    p <- p + ggplot2::geom_smooth(method = trend, se = FALSE)
  }

  p + ggplot2::theme_bw() +
    ggplot2::labs(x = labs[[1]][[1]], y = labs[[1]][[2]],
                  title = labs[[2]][[1]], color = labs[[1]][[3]])
}