Completed #76

NAMESPACE

@@ -10,6 +10,7 @@ export(plot_boxplot)
 export(plot_correlation)
 export(plot_density)
 export(plot_histogram)
+export(plot_intro)
 export(plot_missing)
 export(plot_prcomp)
 export(plot_scatterplot)
@@ -25,7 +26,9 @@ importFrom(networkD3,radialNetwork)
 importFrom(parallel,detectCores)
 importFrom(parallel,mclapply)
 importFrom(rmarkdown,render)
+importFrom(scales,comma)
 importFrom(scales,percent)
+importFrom(stats,complete.cases)
 importFrom(stats,cor)
 importFrom(stats,prcomp)
 importFrom(stats,reorder)

NEWS.md

@@ -1,6 +1,9 @@
 # Changelog
 
 ### DataExplorer 0.6.1.9000
+#### New Features
+* [#76](https://github.com/boxuancui/DataExplorer/issues/76): Added `plot_intro` to visualize results of `introduce`.
+
 #### Enhancements
 * [#42](https://github.com/boxuancui/DataExplorer/issues/42): Applied S3 methods for plotting functions.
 * [#77](https://github.com/boxuancui/DataExplorer/issues/77): `dummify` now works on selected columns.
@@ -17,7 +20,7 @@
 ### DataExplorer 0.6.1
 #### Enhancements
 * Updated vignette for better clarity.
-* Added better error handler for `plot_prcomp`.
+* [#71](https://github.com/boxuancui/DataExplorer/issues/71): Added better error handler for `plot_prcomp`.
 
 #### Bug Fixes
 * [#69](https://github.com/boxuancui/DataExplorer/issues/69): Fixed bug causing `create_report` failure (specifically from `plot_prcomp`) when `y` is specified.

R/introduce.r

@@ -11,10 +11,12 @@
 #'   \item{continuous_columns: number of continuous columns}
 #'   \item{all_missing_columns: number of columns with everything missing}
 #'   \item{total_missing_values: number of missing observations}
+#'   \item{complete_rows: number of rows without missing values. See \link{complete.cases}.}
 #'   \item{total_observations: total number of observations}
-#'   \item{memory_usage: estimated memory allocation in bytes. See \link{object.size}}
+#'   \item{memory_usage: estimated memory allocation in bytes. See \link{object.size}.}
 #' }
 #' @import data.table
+#' @importFrom stats complete.cases
 #' @importFrom utils object.size
 #' @export introduce
 #' @examples
@@ -35,6 +37,7 @@ introduce <- function(data) {
 		"continuous_columns" = split_data[["num_continuous"]],
 		"all_missing_columns" = split_data[["num_all_missing"]],
 		"total_missing_values" = sum(is.na(data)),
+		"complete_rows" = sum(complete.cases(data)),
 		"total_observations" = nrow(data) * ncol(data),
 		"memory_usage" = as.numeric(object.size(data))
 	)

R/plot_intro.r

@@ -0,0 +1,57 @@
+#' Plot introduction
+#'
+#' Plot basic information (from \link{introduce}) for input data.
+#' @param data input data
+#' @param title plot title
+#' @param ggtheme complete ggplot2 themes. The default is \link{theme_gray}.
+#' @param theme_config a list of configurations to be passed to \link{theme}.
+#' @return invisibly return the ggplot object
+#' @keywords plot_intro
+#' @details To change default font family and size, you may pass \code{base_size} and \code{base_family} to \code{ggtheme} options, e.g., \code{ggtheme = theme_gray(base_size = 15, base_family = "serif")}
+#' @details \code{theme_config} argument expects all inputs to be wrapped in a list object, e.g., to change the text color: \code{theme_config = list("text" = element_text(color = "blue"))}
+#' @import ggplot2
+#' @importFrom stats reorder
+#' @importFrom scales comma percent
+#' @export plot_intro
+#' @seealso \link{introduce}
+#' @examples
+#' plot_intro(airquality)
+#' plot_intro(iris)
+
+plot_intro <- function(data, title = NULL, ggtheme = theme_gray(), theme_config = list()) {
+	## Declare variable first to pass R CMD check
+	id <- dimension <- variable <- value <- NULL
+	## Get intro data
+	intro <- introduce(data)
+	memory_usage <- intro[["memory_usage"]]
+	class(memory_usage) <- "object_size"
+	memory_usage_string <- format(memory_usage, unit = "auto")
+	intro2 <- data.table(
+		"id" = seq.int(5L),
+		"dimension" = c(rep("column", 3L), "row", "observation"),
+		"variable" = c("Discrete Columns", "Continuous Columns", "All Missing Columns", "Complete Rows", "Missing Observations"),
+		"value" = c(
+			intro[["discrete_columns"]] / intro[["columns"]],
+			intro[["continuous_columns"]] / intro[["columns"]],
+			intro[["all_missing_columns"]] / intro[["columns"]],
+			intro[["complete_rows"]] / intro[["rows"]],
+			intro[["total_missing_values"]] / intro[["total_observations"]]
+		)
+	)
+	## Create ggplot object
+	output <- ggplot(intro2, aes(x = reorder(variable, -id), y = value, fill = dimension)) +
+		geom_bar(stat = "identity") +
+		geom_text(aes(label = percent(value))) +
+		scale_y_continuous(labels = percent) +
+		scale_fill_discrete("Dimension") +
+		coord_flip() +
+		labs(x = "Metrics", y = "Value")
+	## Plot object
+	class(output) <- c("single", class(output))
+	plotDataExplorer(
+		obj = output,
+		title = ifelse(is.null(title), paste("Memory Usage:", memory_usage_string), title),
+		ggtheme = ggtheme,
+		theme_config = theme_config
+	)
+}

R/plot_missing.r

@@ -23,6 +23,7 @@
 #'
 #' # Plot with customized theme components
 #' plot_missing(airquality, theme_config = list("legend.position" = c("top")))
+
 plot_missing <- function(data, title = NULL, ggtheme = theme_gray(), theme_config = list("legend.position" = c("bottom"))) {
 	## Declare variable first to pass R CMD check
 	pct_missing <- NULL

inst/rmd_template/report.rmd

@@ -54,10 +54,11 @@ if ("introduce" %in% names(report_config)) {
       format(memory_usage, unit = "auto")
     )
   )
-  
   ## Render content
   cat("### Basic Statistics", fill = TRUE)
   kable(intro_df)
+  ## Plot introduction
+  do_call("plot_intro")
 }
 ```
 

vignettes/dataexplorer-intro.Rmd

@@ -104,6 +104,12 @@ introduce(final_data)
 kable(introduce(final_data), format.args = list(big.mark = ","))
 ```
 
+To visualize the table above (with some light analysis):
+
+```{r eda-plot-intro}
+plot_intro(final_data)
+```
+
 ### Missing values
 Real-world data is messy. After running the basic descriptive statistics, you might be interested in the missing data profile. You can simply use `plot_missing` function for this.