distr: Compare Variables with their Distributions

In lares: Analytics & Machine Learning Sidekick

Compare Variables with their Distributions

Description

Compare the distribution of a target variable vs another variable. This function automatically splits into quantiles for numerical variables. Custom and tidyverse friendly.

Usage

distr(
  data,
  ...,
  type = 1,
  ref = TRUE,
  note = NA,
  top = 10,
  breaks = 10,
  na.rm = FALSE,
  force = "none",
  trim = 0,
  clean = FALSE,
  abc = FALSE,
  custom_colours = FALSE,
  plot = TRUE,
  chords = FALSE,
  save = FALSE,
  subdir = NA
)

Arguments

data	Dataframe
...	Variables. Main (target variable) and secondary (values variable) to group by (if needed).
type	Integer. 1 for both plots, 2 for counter plot only, 3 for percentages plot only.
ref	Boolean. Show a reference line if levels = 2? Quite useful when data is unbalanced (not 50/50) because a reference line is drawn.
note	Character. Caption for the plot.
top	Integer. Filter and plot the most n frequent for categorical values.
breaks	Integer. Number of splits for numerical values.
na.rm	Boolean. Ignore NAs if needed.
force	Character. Force class on the values data. Choose between 'none', 'character', 'numeric', 'date'
trim	Integer. Trim labels until the nth character for categorical values (applies for both, target and values)
clean	Boolean. Use cleanText() for categorical values (applies for both, target and values)
abc	Boolean. Do you wish to sort by alphabetical order?
custom_colours	Boolean. Use custom colours function?
plot	Boolean. Return a plot? Otherwise, a table with results
chords	Boolean. Use a chords plot?
save	Boolean. Save the output plot in our working directory
subdir	Character. Into which subdirectory do you wish to save the plot to?

Value

Plot when plot=TRUE with two plots in one: counter distribution grouped by cuts, and proportions distribution grouped by same cuts. data.frame when plot=FALSE with counting, percentages, and cumulative percentages results. When type argument is used, single plots will be returned.

See Also

Other Exploratory: corr_cross(), corr_var(), crosstab(), df_str(), freqs_df(), freqs_list(), freqs_plot(), freqs(), lasso_vars(), missingness(), plot_cats(), plot_df(), plot_nums(), tree_var()

Other Visualization: freqs_df(), freqs_list(), freqs_plot(), freqs(), noPlot(), plot_chord(), plot_survey(), plot_timeline(), tree_var()

Examples

Sys.unsetenv("LARES_FONT") # Temporal
data(dft) # Titanic dataset

# Relation for categorical/categorical values
distr(dft, Survived, Sex)

# Relation for categorical/numeric values
dft %>%
  distr(Survived, Fare, plot = FALSE) %>%
  head(10)
# Sort values
dft %>% distr(Survived, Fare, abc = TRUE)
# Less splits/breaks
dft %>% distr(Survived, Fare, abc = TRUE, breaks = 5)

# Distribution of numerical only
dft[dft$Fare < 20, ] %>% distr(Fare)

# Distribution of numerical/numerical
dft %>% distr(Fare, Age)

# Select only one of the two default plots of distr()
dft %>% distr(Survived, Age, type = 2)
dft %>% distr(Survived, Age, type = 3)

lares documentation built on Nov. 5, 2023, 1:09 a.m.