Title:	Quantile Binned Plots
Version:	0.3.3
Description:	Create quantile binned and conditional plots for Exploratory Data Analysis. The package provides several plotting functions that are all based on quantile binning. The plots are created with 'ggplot2' and 'patchwork' and can be further adjusted.
License:	MIT + file LICENSE
URL:	https://edwindj.github.io/qbinplots/
BugReports:	https://github.com/edwindj/qbinplots/issues
Encoding:	UTF-8
RoxygenNote:	7.3.2
Depends:	R (≥ 4.1.0)
Imports:	ggplot2, data.table, patchwork, scales
Suggests:	palmerpenguins, tinytest
NeedsCompilation:	no
Packaged:	2025-02-22 10:22:13 UTC; edwin
Author:	Edwin de Jonge [aut, cre], Martijn Tennekes [ctb]
Maintainer:	Edwin de Jonge <edwindjonge@gmail.com>
Repository:	CRAN
Date/Publication:	2025-02-24 17:20:02 UTC

qbinplots

Description

This package creates plots using quantile binning.

Details

Quantile binning is an exploratory data analysis tool that helps to see the distribution of the variables in a dataset as a function of the variable that is binned.

A data.frame is quantile binned on a variable x using qbin() and then plotted with one of the avaible plot functions.

qbinplots offers various types of plots:

• ⁠qbin_*⁠ quantile binned plots that show the distribution of the variables in the quantile bins.

• ⁠cond_*⁠ conditional quantile plots that show the distribution of the variables conditional on the x variable.

Quantile binned plots

• qbin_lineplot() highlights the change in median between qbins, shows the distribution within qbins.

• qbin_barplot() shows the size of medians or expected value of qbins.

• qbin_boxplot() shows the distribution within qbins.

• qbin_heatmap() shows the distribution within the qbins.

Conditional (quantile binned) plots

• cond_boxplot() shows the distribution of the variables conditional on the x variable.

• cond_barplot() shows the expected median/mean of the variables conditional on the x variable.

• funq_plot() shows a functional view of the data, plotting the median and interquartile range of numerical variables and level frequency of the other variables as a function of the x variable using quantile bins.

Author(s)

Maintainer: Edwin de Jonge edwindjonge@gmail.com (ORCID)

Other contributors:

• Martijn Tennekes mtennekes@gmail.com [contributor]

See Also

Useful links:

• https://edwindj.github.io/qbinplots/

• Report bugs at https://github.com/edwindj/qbinplots/issues

Conditional quantile barplot

Description

cond_barplot() conditions all variables on x by quantile binning and shows the median or mean of the other variables for each x.

Usage

cond_barplot(
  data,
  x = NULL,
  n = 100,
  min_bin_size = NULL,
  overlap = NULL,
  ncols = NULL,
  fill = "#2f4f4f",
  auto_fill = FALSE,
  show_bins = FALSE,
  type = c("median", "mean"),
  ...
)

Arguments

Value

A list of ggplot objects.

See Also

Other conditional quantile plotting functions: cond_boxplot(), cond_heatmap(), funq_plot()

Examples

# plots the expected median conditional on Sepal.Width
cond_barplot(iris, "Sepal.Width", n = 12)



  # plots the expected median
  cond_barplot(iris, "Sepal.Width", n = 12, show_bins = TRUE)

  data("diamonds", package="ggplot2")

  cond_barplot(diamonds[c(1:4, 7)], "carat", auto_fill = TRUE)

  if (require(palmerpenguins)) {
    p <- cond_barplot(penguins[1:7], "body_mass_g", auto_fill = TRUE)
    print(p)

    # compare with qbin_boxplot
    p <- cond_boxplot(penguins[1:7], "body_mass_g", auto_fill = TRUE)
    print(p)
  }

Conditional quantile boxplot

Description

cond_boxplot() conditions all variables on x by quantile binning and shows the boxplots for the other variables for each value of qbinned x.

Usage

cond_boxplot(
  data,
  x = NULL,
  n = 100,
  min_bin_size = NULL,
  color = "#002f2f",
  fill = "#2f4f4f",
  auto_fill = FALSE,
  ncols = NULL,
  xmarker = NULL,
  qmarker = NULL,
  show_bins = FALSE,
  xlim = NULL,
  connect = FALSE,
  ...
)

Arguments

Details

cond_boxplot is the same function as funq_plot() but with different defaults, namely connect = FALSE and auto_fill = FALSE. funq_plot highlights the functional relationship between x and the y-variables, by connecting the medians of the quantile bins.

qbin_boxplot() shows the boxplots of the quantile bins on a quantile scale.

Value

A list of ggplot objects.

See Also

Other conditional quantile plotting functions: cond_barplot(), cond_heatmap(), funq_plot()

Examples

cond_boxplot(
  iris, x = "Petal.Length"
)

Conditional heatmap

Description

cond_heatmap shows the conditional distribution of the y of variables for each quantile bin of x. It is an alternative to cond_boxplot(), fine graining the distribution per qbin(). cond_barplot() highlights the median/mean of the quantile bins, while funq_plot() highlights the functional dependency of the median.

Usage

cond_heatmap(
  data,
  x = NULL,
  n = 100,
  min_bin_size = NULL,
  overlap = NULL,
  bins = c(n, 25),
  ncols = NULL,
  auto_fill = FALSE,
  show_bins = FALSE,
  fill = "#2f4f4f",
  low = "#eeeeee",
  high = "#2f4f4f",
  ...
)

Arguments

Value

A list of ggplot objects.

See Also

Other conditional quantile plotting functions: cond_barplot(), cond_boxplot(), funq_plot()

Examples

cond_heatmap(
  iris,
  x = "Petal.Length",
  n = 12
)



  data("diamonds", package="ggplot2")

  cond_heatmap(
    diamonds,
    x = "carat",
    bins <- c(100,100)
  )[6:8]

Functional quantile plot

Description

funq_plot() conditions on variable x with quantile binning and plots the median and interquartile range of numerical variables and level frequency of the other variables as a function the x variable.

Usage

funq_plot(
  data,
  x = NULL,
  n = 100,
  min_bin_size = NULL,
  overlap = NULL,
  color = "#002f2f",
  fill = "#2f4f4f",
  auto_fill = TRUE,
  ncols = NULL,
  xmarker = NULL,
  qmarker = NULL,
  show_bins = FALSE,
  xlim = NULL,
  connect = TRUE,
  ...
)

Arguments

Details

By highlighting and connecting the median values it creates a functional view of the data. What is the (expected) median given a certain value of x?

It qbins the x variable and plots the medians of the qbins vs the other variables, thereby creating a functional view of x to the rest of the data, calculating the statistics for each bin, hence the name funq_plot.

Value

A ggplot object with the plots

See Also

Other conditional quantile plotting functions: cond_barplot(), cond_boxplot(), cond_heatmap()

Examples

funq_plot(iris, "Sepal.Length", xmarker=5.5)



  funq_plot(
    iris,
    x = "Sepal.Length",
    xmarker=5.5,
    overlap = TRUE
  )


  data("diamonds", package="ggplot2")
  funq_plot(diamonds[1:7], "carat", xlim=c(0,2))

  if (require(palmerpenguins)){
    funq_plot(
      penguins[1:7],
      x = "body_mass_g",
      xmarker=4650,
      ncol = 3
    )
  }

Bin a data.frame into quantile bins

Description

Bins a data.frame into quantile bins for variable x in data.

Usage

qbin(data, x = NULL, n = 100, min_bin_size = NULL, overlap = NULL, ...)

Arguments

Details

Each numeric variable in the data.frame is binned into n quantile bins, for which the fivenum() and mean() is calculated.

When n/nrow(data) is less than min_bin_size, qbin gives a warning and n is adjusted to nrow(data)/min_bin_size. Each categorical variable is binned into n quantile bins, for which the level frequency is calculated.

Value

a qbin object with:

• $x the variable name used for binning

• $bin a vector of bin numbers

• $n the number of bins

• $num_cols a vector of numeric column names

• $cat_cols a vector of categorical column names

• $data a list of data.tables with the collected information

Quantile binned bar plot

Description

qbin_barplot() shows the median or mean for each quantile bin, thereby focusing on the expected value per qbin(). For a conditional plot, see cond_barplot().

Usage

qbin_barplot(
  data,
  x = NULL,
  n = 100,
  min_bin_size = NULL,
  overlap = NULL,
  ncols = NULL,
  fill = "#2f4f4f",
  type = c("median", "mean"),
  ...
)

table_plot(data, x = NULL, n = 100, ncols = ncol(data), fill = "#555555", ...)

Arguments

Details

The table_plot is a specific form of qbin_barplot with ncols set to ncol(data).

Value

A list of ggplot objects.

See Also

Other qbin plotting functions: qbin_boxplot(), qbin_heatmap(), qbin_lineplot()

Examples

  data("diamonds", package="ggplot2")

  table_plot(diamonds[c(1:4, 7)], "carat")

  qbin_barplot(iris, "Sepal.Length", n = 12)

  table_plot(iris, "Sepal.Length", n=12)
  table_plot(
    iris,
    x = "Sepal.Length",
    min_bin_size=20,
    overlap=TRUE
  )

  if (require(palmerpenguins)) {
    table_plot(penguins[1:7], "body_mass_g", 19)
  }

Quantile binned boxplot

Description

qbin_boxplot creates quantile binned boxplots from data using x as the binning variable. It focuses on the change of median between qbins. It is a complement to qbin_heatmap() which focuses on the distribution within the qbins.

Usage

qbin_boxplot(
  data,
  x = NULL,
  n = 100,
  min_bin_size = NULL,
  ncols = NULL,
  overlap = NULL,
  connect = FALSE,
  color = "#002f2f",
  fill = "#2f4f4f",
  auto_fill = FALSE,
  qmarker = NULL,
  xmarker = NULL,
  ...
)

Arguments

Details

The data is binned by the x and a boxplot is created for each bin. The median of the subsequent boxplots are connected to highlight jumps in the data. It hints at the dependecy of the variable on the binning variable.

Value

A list of ggplot objects.

See Also

Other qbin plotting functions: qbin_barplot(), qbin_heatmap(), qbin_lineplot()

Examples

qbin_boxplot(
  iris,
  x = "Sepal.Length",
)


  qbin_boxplot(
    iris,
    x = "Sepal.Length",
    connect = TRUE,
    overlap = TRUE
  )

  qbin_boxplot(
    iris,
    x = "Sepal.Length",
    connect = TRUE,
    xmarker = 5.5,
    auto_fill = TRUE
  )

  data("diamonds", package="ggplot2")

  qbin_boxplot(
    diamonds[1:7],
    "carat",
    auto_fill = TRUE
  )

  qbin_boxplot(
    diamonds[1:7],
    "price",
    auto_fill = TRUE,
  )

Quantile binned heatmap

Description

qbin_heatmap shows the distribution of the y of variables for each quantile bin of x. It is an alternative to qbin_boxplot(), fine graining the distribution per qbin(). qbin_barplot() highlights the median/mean of the quantile bins, while

Usage

qbin_heatmap(
  data,
  x = NULL,
  n = 25,
  min_bin_size = NULL,
  overlap = NULL,
  bins = c(n),
  type = c("gradient", "size"),
  ncols = NULL,
  auto_fill = FALSE,
  fill = "#2f4f4f",
  low = "#eeeeee",
  high = "#2f4f4f",
  ...
)

Arguments

Value

A list of ggplot objects.

See Also

Other qbin plotting functions: qbin_barplot(), qbin_boxplot(), qbin_lineplot()

Examples

  qbin_heatmap(
    iris,
    "Sepal.Length",
    auto_fill = TRUE
  )

  qbin_heatmap(
    iris,
    "Sepal.Length",
    auto_fill = TRUE,
    type = "size"
  )

  qbin_heatmap(
    iris,
    "Sepal.Length",
    overlap = TRUE,
    auto_fill = TRUE
  )

  data("diamonds", package="ggplot2")

  qbin_heatmap(
    diamonds[c(1,7:9)],
    x = "price",
    n = 150
  )

Quantile binned lineplot

Description

qbin_lineplot creates quantile binned boxplots from data using x as the binning variable and connects the medians: it focuses on the change of median between qbins.

Usage

qbin_lineplot(
  data,
  x = NULL,
  n = 100,
  min_bin_size = NULL,
  ncols = NULL,
  connect = TRUE,
  color = "#002f2f",
  fill = "#2f4f4f",
  auto_fill = FALSE,
  qmarker = NULL,
  xmarker = NULL,
  ...
)

Arguments

Details

The data is binned by the x and a boxplot is created for each bin. The median of the subsequent boxplots are connected to highlight jumps in the data. It hints at the dependecy of the variable on the binning variable.

Value

A list of ggplot objects.

See Also

Other qbin plotting functions: qbin_barplot(), qbin_boxplot(), qbin_heatmap()

Examples

qbin_lineplot(
  iris,
  x = "Sepal.Length",
)


  qbin_lineplot(
    iris,
    x = "Sepal.Length",
    xmarker = 5.5,
    auto_fill = TRUE
  )

  qbin_lineplot(
    iris,
    x = "Sepal.Length",
    overlap=TRUE,
    xmarker = 5.5,
    auto_fill = TRUE
  )

  data("diamonds", package="ggplot2")

  qbin_lineplot(
    diamonds[1:7],
    "carat",
    auto_fill = TRUE
  )

  qbin_lineplot(
    diamonds[1:7],
    "price",
    auto_fill = TRUE,
  )