describe_df: Describe and visulaize the statistic distribution of a...

In madlogos/aseskit: ASES Generic Toolkit

Description

A family of functions to analyze a data.frame and output the descriptive analysis results and/or visualization.

Usage

describe_df(data, qual_vars, quan_vars, nrow = 1, ncol = 1, ...)

cal_df_distrib(data, qual_vars, quan_vars, ...)

vis_df_distrib(data, qual_vars, quan_vars, nrow = NULL, ncol = NULL,
  ...)

Arguments

data	a data.frame
qual_vars	variable(s) indicating the qualitative variable(s). It accepts five forms: missing (default), the function will automatically identify the qualitative variable character vector indicating variable names, e.g., c("cyl", "gear") integer vector indicating variable index, e.g., c(2, 10) quosures yielded using vars(), e.g, vars(cyl, gear) NULL indicating that the function will not process qualitatitive variables
quan_vars	variable(s) indicating the quantitative variable(s). It accepts five forms: missing (default), the function will automatically identify the quantitative variables character vector indicating variable names, e.g., c("drat", "mpg") integer vector indicating variable index, e.g., c(1, 5) quosures yielded using vars(), e.g, vars(drat, mpg) NULL indicating that the function will not process quantitative variables
nrow	NULL or positive integer. Only applicable for vis_df_distrib, determining the number of rows in the plot grid. See facet_wrap for more details.
ncol	NULL or positive integer. Only applicable for vis_df_distrib, determining the number of columns in the plot grid. See facet_wrap for more details.
...	other arguments to pass to describe_df, cal_df_distrib and vis_df_distrib. Accepts the following: pass to cal_df_distrib argument for choose_col_idx diversity_threshold pass to vis_df_distrib help_on_dots logical. If TRUE, you can get help info for ... argument. elipsis arguments for geom_histogram hist.title, hist.tag, hist.subtitle, hist.stat, hist.position, hist.bins, hist.color, hist.inherit.aes elipsis arguments for histograms and geom_bar bar.stat, bar.position, bar.width, bar.binwidth, bar.fill, bar.color, bar.size, bar.alpha, bar.na.rm, bar.show.legend, bar.inheirt.aes elipsis arguments for geom_vline vline.color, vline.size, vline.alpha, vline.show.legend, vline.na.rm

Details

describe_df is a wrapper of cal_df_distrib and vis_df_distrib.

• cal_df_distrib, built on top of describe_num_cols and describe_chr_cols, returns two lists: one for quantitative variables and the other for qualitative ones.
  

• vis_df_distrib, built on top of vis_num_cols and vis_chr_cols, returns two sets of plots: one for quantitative variables and the other for qualitative ones, each of which contains one or multiple ggplots objects.

Value

describe_df(): a list of results by cal_df_distrib and vis_df_distrib

cal_df_distrib(): two lists of tibbles

'qual'

a tibble with 4 columns: <varname>, 'value', 'freq', 'prop'; or a character "no match columns"

'quan'

a tibble with 11 columns: <varname>, 'count', 'n_na', 'p_na', 'mean', 'sd', 'min', 'lower', 'median', 'higher', 'max'; or a character "no match columns"

vis_df_distrib(): two sets of plots

• nrow and ncol are 1 respectively, then the distribution plots (histogram or bar plots) will be shown one by one.
  

• either nrow or ncol is more than 1, then the plots will be organized in facetted plots (with facet_wrap()).
  

• If both nrow and ncol are NULL, then the plots will be organized into one plot

Examples

## Not run: 
describe_df(mtcars, vars(am, cyl, mpg, wt), NULL)

# you can also detect the distribution of the data in clipboard
describe_df(parse_clipb())

## End(Not run)
## Not run: 
cal_df_distrib(iris)  ## or
library(dplyr)
cal_df_distrib(iris, quan_vars=vars(Sepal.Length, Petal.Length))

## End(Not run)
## Not run: 
library(dplyr)
vis_df_distrib(mtcars, qual=vars(am, cyl), quan=vars(mpg, wt))

## End(Not run)

madlogos/aseskit documentation built on June 26, 2019, 12:17 a.m.