Title: Turnkey Visualisations for Exploratory Data Analysis
Version: 0.1.0
Description: Provides interactive visualisations for exploratory data analysis of high-dimensional datasets. Includes parallel coordinate plots for exploring large datasets with mostly quantitative features, but also stacked one-dimensional visualisations that more effectively show missingness and complex categorical relationships in smaller datasets.
License: MIT + file LICENSE
Encoding: UTF-8
RoxygenNote: 7.3.2
URL: https://github.com/CCICB/ggEDA, https://ccicb.github.io/ggEDA/
BugReports: https://github.com/CCICB/ggEDA/issues
Imports: assertions (≥ 0.2.0), cli, ggiraph (≥ 0.8.11), ggplot2, ggtext, grDevices, patchwork (≥ 1.3.0), rank (≥ 0.1.1), rlang
Suggests: covr, infotheo, knitr, rmarkdown, testthat (≥ 3.0.0), TSP
Config/Needs/website: uwot, datarium, palmerpenguins
Config/testthat/edition: 3
Depends: R (≥ 3.5)
LazyData: true
NeedsCompilation: no
Packaged: 2025-05-05 07:54:36 UTC; selkamand
Author: Sam El-Kamand ORCID iD [aut, cre], Children's Cancer Institute Australia [cph]
Maintainer: Sam El-Kamand <sam.elkamand@gmail.com>
Repository: CRAN
Date/Publication: 2025-05-07 12:00:02 UTC

ggEDA: Turnkey Visualisations for Exploratory Data Analysis

Description

logo

Provides interactive visualisations for exploratory data analysis of high-dimensional datasets. Includes parallel coordinate plots for exploring large datasets with mostly quantitative features, but also stacked one-dimensional visualisations that more effectively show missingness and complex categorical relationships in smaller datasets.

Author(s)

Maintainer: Sam El-Kamand sam.elkamand@gmail.com (ORCID)

Other contributors:

See Also

Useful links:

Baseball Fans Dataset

Description

An artificially generated dataset describing basic demographics and accessorization choices of baseball fans as part of a a hypothetical market research study from stadium merchandise vendors. None of the data are real; they were produced for illustrative and testing purposes only.

Usage

baseballfans

Format

baseballfans

A data frame with 19 rows and 10 columns:

ID

Unique integer identifier for each individual.

Age

Age in years at time of observation.

Gender

Self‐reported gender (“Male” or “Female”).

EyeColour

Eye color (“Brown”, “Green”, “Blue”), or missing (NA) if not recorded.

Height

Height in centimeters; missing (NA) if not recorded.

HairColour

Hair color (“Black”, “Blond”, “Red”, “Brown”).

Glasses

Logical flag (TRUE/FALSE) indicating whether the individual wears glasses.

WearingHat

Logical flag (TRUE/FALSE) indicating whether the individual is wearing a hat.

WearingHat_tooltip

Type of hat worn, if any (e.g., “baseball cap”, “stetson”, “fedora”, “top hat”); empty when WearingHat == FALSE.

Date

Date of observation in day/month/year format (e.g., 9/05/2023). Stored as character vector

#' @source Synthetic data; no real persons were observed.

Details

This mock dataset was created to demonstrate ggEDA functionality. All entries are fictional.

Make strings prettier for printing

Description

Takes an input string and 'beautify' by converting underscores to spaces and

Usage

beautify(string, autodetect_units = TRUE)

Arguments

string

input string

autodetect_units

automatically detect units (e.g. mm, kg, etc) and wrap in brackets.

Value

string

Parse a tibble and ensure it meets standards

Description

Parse a tibble and ensure it meets standards

Usage

column_info_table(
  data,
  maxlevels = 6,
  col_id = NULL,
  cols_to_plot,
  tooltip_column_suffix = "_tooltip",
  ignore_column_regex = "_ignore$",
  palettes,
  colours_default,
  colours_default_logical,
  verbose
)

Arguments

data

data.frame to autoplot (data.frame)

maxlevels

for categorical variables, what is the maximum number of distinct values to allow (too many will make it hard to find a palette that suits). (number)

col_id

name of column to use as an identifier. If null, artificial IDs will be created based on row-number.

cols_to_plot

names of columns in data that should be plotted. By default plots all valid columns (character)

tooltip_column_suffix

the suffix added to a column name that indicates column should be used as a tooltip (string)

ignore_column_regex

a regex string that, if matches a column name, will cause that column to be excluded from plotting (string). If NULL no regex check will be performed. (default: "_ignore$")

palettes

A list of named vectors. List names correspond to data column names (categorical only). Vector names to levels of columns. Vector values are colours, the vector names are used to map values in data to a colour.

colours_default

Default colors for categorical variables without a custom palette.

colours_default_logical

Colors for binary variables: a vector of three colors representing TRUE, FALSE, and NA respectively (character).

verbose

Numeric value indicating the verbosity level:

  • 2: Highly verbose, all messages.

  • 1: Key messages only.

  • 0: Silent, no messages.

Value

tibble with the following columns:

  1. colnames

  2. coltype (categorical/numeric/tooltip/invalid)

  3. ndistinct (number of distinct values)

  4. plottable (should this column be plotted)

  5. tooltip_col (the name of the column to use as the tooltip) or NA if no obvious tooltip column found

Count Edge Crossings for All Numeric Column Pairs

Description

Computes the total number of edge crossings between all pairs of numeric columns in a given dataset.

Usage

count_all_edge_crossings(
  data,
  approximate = FALSE,
  subsample_prop = 0.4,
  recalibrate = FALSE
)

Arguments

data

A data.frame or tibble containing the dataset. Only numeric columns are considered for edge crossing calculations.

approximate

estimate crossings based on a subsample of the data. See subsample_prop for details.

subsample_prop

only used when approximate = TRUE. If 0-1, controls the proportion of data to be sampled to speed up computation. If a whole number other than 0 or 1, represents the number of rows subsampled

recalibrate

when approximating crossings via subsetting, should number of crossings calculated for the subsample be upscaled to match the full count. (turned off by default since it amplifies sampling error).

Details

The function:

  1. Filters the input data to retain only numeric columns.

  2. Computes all possible pairs of numeric columns.

  3. Uses count_edge_crossings() to calculate crossings for each pair.

  4. Returns the results in a summarized data frame.

Value

A data.frame with three columns:

col1

The name of the first column in the pair.

col2

The name of the second column in the pair.

crossings

Total number of edge crossings for that pair.

Count Edge Crossings in Parallel Coordinates

Description

Calculates the total number of edge crossings between two numeric vectors in a 2-column parallel coordinates setup. Each axis represents one of the columns.

Usage

count_edge_crossings(l, r)

Arguments

l

A numeric vector representing values on the left axis. Must have the same length as r.

r

A numeric vector representing values on the right axis. Must have the same length as l.

Details

An edge crossing occurs when two edges intersect between the axes. Formally, edges (l[i], r[i]) and (l[j], r[j]) cross if (l[i] - l[j]) * (r[i] - r[j]) < 0.

Value

An integer indicating the total number of edge crossings.

Create a Distance Matrix from Edge Crossing Data

Description

Converts the results of count_all_edge_crossings() into a distance matrix, where each entry represents the number of crossings between two columns.

Usage

create_distance_matrix(data, as.dist = FALSE)

Arguments

data

A data frame with columns col1, col2, and crossings.

as.dist

Logical; if TRUE, converts the matrix to a dist object.

Value

A square matrix of distances, or a dist object if as.dist = TRUE.

Reorder Factor Levels by Descending Frequency

Description

Reorders the levels of a factor by their frequency, in descending order.

Usage

fct_infreq(x)

Arguments

x

A factor or an object coerced to a factor.

Value

A factor with levels ordered by descending frequency.

Relevel Factor by Specified Levels

Description

Reorder the levels of a factor by moving specified levels to a new position.

Usage

fct_relevel_base(x, ..., after = 0)

Arguments

x

A factor to be releveled.

...

Levels to move in the factor.

after

A numeric scalar specifying the position after which the moved levels should be placed. Use 0 to place them at the front.

Value

A factor with the specified levels moved to the chosen position.

Reverse the Levels of a Factor

Description

Reverses the existing level order of a factor.

Usage

fct_rev(x)

Arguments

x

A factor or an object coerced to a factor.

Value

A factor with reversed levels.

Compute the Total Path Distance for an Axis Order

Description

Given a sequence of axis names and a distance matrix, sums pairwise distances along the path.

Usage

feature_vector_to_total_path_distance(axis_names, mx)

Arguments

axis_names

A character vector indicating the axis order.

mx

A matrix of distances, with row and column names matching axis_names.

Value

A numeric value representing the total distance.

Optimize Axis Ordering Directly from a Data Frame

Description

Computes the number of edge crossings between all numeric columns in data, converts this information into a distance matrix, and then determines an optimal ordering of the columns based on the specified method.

Usage

get_optimal_axis_order(
  data,
  verbose = TRUE,
  method = "auto",
  metric = c("mutinfo", "crossings", "crossings_fast"),
  return_detailed = FALSE
)

Arguments

data

A data.frame or tibble containing the dataset. Only numeric columns are considered for edge crossing calculations.

verbose

A logical value; if TRUE, prints progress messages.

method

A character string specifying the method. Options are "auto", "brute_force", or "repetitive_nn_with_2opt".

metric

which metric should take as the distance between axes to minimise. mutual information: minimise mutual distance (1- uniminmax of mutinfo similarity matrix calculated by emp) crossings: minimise the total number of edge crossings (warning: slow to compute for large datasets). crossings_fast: same as above but calculates crossings on a subset of data (100 rows)

return_detailed

A logical; if TRUE, returns a list with additional data (e.g., intermediate calculations) for debugging.

Value

A character vector of axis names in the chosen order, or a list with additional data if return_detailed = TRUE.

Parallel Coordinate Plots

Description

Visualize relationships between numeric variables and categorical groupings using parallel coordinate plots.

Usage

ggparallel(
  data,
  col_id = NULL,
  col_colour = NULL,
  highlight = NULL,
  interactive = TRUE,
  order_columns_by = c("appearance", "random", "auto"),
  order_observations_by = c("frequency", "original"),
  verbose = TRUE,
  palette_colour = palette.colors(palette = "Set2"),
  palette_highlight = c("red", "grey90"),
  convert_binary_numeric_to_factor = TRUE,
  scaling = c("uniminmax", "none"),
  return = c("plot", "data"),
  options = ggparallel_options()
)

Arguments

data

A data frame containing the variables to plot.

col_id

The name of the column to use as an identifier. If NULL, artificial IDs will be generated based on row numbers. (character)

col_colour

Name of the column to use for coloring lines in the plot. If NULL, no coloring is applied. (character)

highlight

A level from col_colour to emphasize in the plot. Ignored if col_colour is not set. (character)

interactive

Produce interactive ggiraph visualiastion (flag)

order_columns_by

Strategy for ordering columns in the plot. Options include:

  • "appearance": Order columns by their order in data (default).

  • "random": Randomly order columns.

  • "auto": Automatically order columns based on context:

    • If highlight is set, columns are ordered to maximize separation between the highlighted level and all others, using mutual information.

    • If col_colour is set but highlight is not, columns are ordered based on mutual information with all classes in col_colour.

    • If neither highlight nor col_colour is set, columns are ordered to minimize the estimated number of crossings, using a repetitive nearest neighbour approach with two-opt refinement.

order_observations_by

Strategy for ordering lines in the plot. Options include:

  • "frequency": Draw the largest groups first.

  • "original": Preserve the original order in data.

Ignored if highlight is set.

verbose

Logical; whether to display informative messages during execution. (default: TRUE)

palette_colour

A named vector of colors for categorical levels in col_colour. (default: Set2 palette)

palette_highlight

A two-color vector for highlighting (highlight and others). (default: c("red", "grey90"))

convert_binary_numeric_to_factor

Logical; whether to convert numeric columns containing only 0, 1, and NA to factors. (default: TRUE)

scaling

Method for scaling numeric variables. Options include:

  • "uniminmax": Rescale each variable to range [0, 1].

  • "none": No rescaling. Use raw values.

return

What to return. Options include:

  • "plot": Return the ggplot object (default).

  • "data": Return the processed data used for plotting.

options

A list of additional visualization parameters created by ggparallel_options().

Value

A ggplot object or a processed data frame, depending on the return parameter.

Examples

ggparallel(
  data = minibeans,
  col_colour = "Class",
  order_columns_by = "auto"
)

ggparallel(
  data = minibeans,
  col_colour = "Class",
  highlight = "DERMASON",
  order_columns_by = "auto"
)

# Customise appearance using options argument
ggparallel(
  data = minibeans,
  col_colour = "Class",
  order_columns_by = "auto",
  options = ggparallel_options(show_legend = FALSE)
)

Visual Parameters for ggparallel Plots

Description

Configures aesthetic and layout settings for plots generated by ggparallel.

Usage

ggparallel_options(
  show_legend = TRUE,
  show_legend_titles = FALSE,
  legend_position = c("bottom", "right", "left", "top"),
  legend_title_position = c("left", "top", "bottom", "right"),
  legend_nrow = NULL,
  legend_ncol = NULL,
  legend_key_size = 1,
  beautify_text = TRUE,
  max_digits_bounds = 1,
  x_axis_text_angle = 90,
  x_axis_text_hjust = 0,
  x_axis_text_vjust = 0.5,
  fontsize_x_axis_text = 12,
  show_column_names = TRUE,
  show_points = FALSE,
  show_bounds_labels = FALSE,
  show_bounds_rect = FALSE,
  line_alpha = 0.5,
  line_width = NULL,
  line_type = 1,
  x_axis_gridlines = ggplot2::element_line(colour = "black"),
  interactive_svg_width = NULL,
  interactive_svg_height = NULL
)

Arguments

show_legend

Display the legend on the plot (flag).

show_legend_titles

Display titles for legends (flag).

legend_position

Position of the legend ("right", "left", "bottom", "top").

legend_title_position

Position of the legend title ("top", "bottom", "left", "right").

legend_nrow

Number of rows in the legend (number).

legend_ncol

Number of columns in the legend. If set, legend_nrow should be NULL (number).

legend_key_size

Size of the legend key symbols. (number).

beautify_text

Beautify y-axis text and legend titles by capitalizing words and adding spaces (flag).

max_digits_bounds

Number of digits to round the axis bounds label text to (number)

x_axis_text_angle

Angle of the x axis text describing column names (number)

x_axis_text_hjust

Horizontal Justification of the x axis text describing column names (number)

x_axis_text_vjust

Vertical Justification of the x axis text describing column names (number)

fontsize_x_axis_text

fontsize of the x-axis text describing column names (number)

show_column_names

Show column names as x axis text (flag)

show_points

Show points (flag)

show_bounds_labels

Show bounds (min and max value) of each feature with labels above / below the axes (flag)

show_bounds_rect

Show bounds (min and max value) of each feature with a rectangular graphic (flag)

line_alpha

Alpha of line geom (number)

line_width

Width of the line geom (number)

line_type

Type of line geom (number or string. see ggplot2::aes_linetype_size_shape() for valid options)

x_axis_gridlines

Customise look of x axis gridlines. Must be either a call to ggplot2::element_line() or ggplot2::element_blank().

interactive_svg_width, interactive_svg_height

Width and height of the interactive graphic region (in inches). Only used when interactive = TRUE.

Value

A list of visualization parameters for ggparallel.

Examples

ggparallel(
  data = minibeans,
  col_colour = "Class",
  order_columns_by = "auto"
)

ggparallel(
  data = minibeans,
  col_colour = "Class",
  highlight = "DERMASON",
  order_columns_by = "auto"
)

# Customise appearance using options argument
ggparallel(
  data = minibeans,
  col_colour = "Class",
  order_columns_by = "auto",
  options = ggparallel_options(show_legend = FALSE)
)

AutoPlot an entire data.frame

Description

Visualize all columns in a data frame with ggEDA's vertically aligned plots and automatic plot selection based on variable type. Plots are fully interactive, and custom tooltips can be added.

Usage

ggstack(
  data,
  col_id = NULL,
  col_sort = NULL,
  order_matches_sort = TRUE,
  maxlevels = 7,
  verbose = 2,
  drop_unused_id_levels = FALSE,
  interactive = TRUE,
  return = c("plot", "column_info", "data"),
  palettes = NULL,
  sort_type = c("frequency", "alphabetical"),
  desc = TRUE,
  limit_plots = TRUE,
  max_plottable_cols = 10,
  cols_to_plot = NULL,
  tooltip_column_suffix = "_tooltip",
  ignore_column_regex = "_ignore$",
  convert_binary_numeric_to_factor = TRUE,
  options = ggstack_options(show_legend = !interactive)
)

Arguments

data

data.frame to autoplot (data.frame)

col_id

name of column to use as an identifier. If null, artificial IDs will be created based on row-number.

col_sort

name of columns to sort on. To do a hierarchical sort, supply a vector of column names in the order they should be sorted (character).

order_matches_sort

should the column plots be stacked top-to-bottom in the order they appear in col_sort (flag)

maxlevels

for categorical variables, what is the maximum number of distinct values to allow (too many will make it hard to find a palette that suits). (number)

verbose

Numeric value indicating the verbosity level:

  • 2: Highly verbose, all messages.

  • 1: Key messages only.

  • 0: Silent, no messages.

drop_unused_id_levels

if col_id is a factor with unused levels, should these be dropped or included in visualisation

interactive

produce interactive ggiraph visualiastion (flag)

return

a string describing what this function should return. Options include:

  • plot: Return the ggEDA visualisation (default)

  • colum_info: Return a data.frame describing the columns the dataset.

  • data: Return the processed dataset used for plotting.

palettes

A list of named vectors. List names correspond to data column names (categorical only). Vector names to levels of columns. Vector values are colours, the vector names are used to map values in data to a colour.

sort_type

controls how categorical variables are sorted. Numerical variables are always sorted in numerical order irrespective of the value given here. Options are alphabetical or frequency

desc

sort in descending order (flag)

limit_plots

throw an error when there are > max_plottable_cols in dataset (flag)

max_plottable_cols

maximum number of columns that can be plotted (default: 10) (number)

cols_to_plot

names of columns in data that should be plotted. By default plots all valid columns (character)

tooltip_column_suffix

the suffix added to a column name that indicates column should be used as a tooltip (string)

ignore_column_regex

a regex string that, if matches a column name, will cause that column to be excluded from plotting (string). If NULL no regex check will be performed. (default: "_ignore$")

convert_binary_numeric_to_factor

If a numeric column conatins only values 0, 1, & NA, then automatically convert to a factor.

options

a list of additional visual parameters created by calling ggstack_options(). See ggstack_options for details.

Value

ggiraph interactive visualisation

Examples


# Create Basic Plot
ggstack(baseballfans, col_id = "ID", col_sort = "Glasses")

# Configure plot ggstack_options()
ggstack(
  lazy_birdwatcher,
  col_sort = "Magpies",
  palettes = list(
    Birdwatcher = c(Robert = "#E69F00", Catherine = "#999999"),
    Day = c(Weekday = "#999999", Weekend = "#009E73")
  ),
  options = ggstack_options(
    show_legend = TRUE,
    fontsize_barplot_y_numbers = 12,
    legend_text_size = 16,
    legend_key_size = 1,
    legend_nrow = 1,
  )
)

Visual Parameters for ggstack Plots

Description

Configures aesthetic and layout settings for plots generated by ggstack.

Usage

ggstack_options(
  colours_default = c("#66C2A5", "#FC8D62", "#8DA0CB", "#E78AC3", "#A6D854", "#FFD92F",
    "#E5C494"),
  colours_default_logical = c(`TRUE` = "#648fff", `FALSE` = "#dc267f"),
  colours_missing = "grey90",
  show_legend_titles = FALSE,
  legend_title_position = c("top", "bottom", "left", "right"),
  legend_nrow = 4,
  legend_ncol = NULL,
  legend_title_size = NULL,
  legend_text_size = NULL,
  legend_key_size = 0.3,
  legend_orientation_heatmap = c("horizontal", "vertical"),
  show_legend = TRUE,
  legend_position = c("right", "left", "bottom", "top"),
  na_marker = "!",
  na_marker_size = 8,
  na_marker_colour = "black",
  show_na_marker_categorical = FALSE,
  show_na_marker_heatmap = FALSE,
  colours_heatmap_low = "purple",
  colours_heatmap_high = "seagreen",
  transform_heatmap = c("identity", "log10", "log2"),
  fontsize_values_heatmap = 3,
  show_values_heatmap = FALSE,
  colours_values_heatmap = "white",
  vertical_spacing = 0,
  numeric_plot_type = c("bar", "heatmap"),
  y_axis_position = c("left", "right"),
  width = 0.9,
  relative_height_numeric = 4,
  cli_header = "Running ggstack",
  interactive_svg_width = NULL,
  interactive_svg_height = NULL,
  fontsize_barplot_y_numbers = 8,
  max_digits_barplot_y_numbers = 3,
  fontsize_y_title = 12,
  beautify_text = TRUE
)

Arguments

colours_default

Default colors for categorical variables without a custom palette.

colours_default_logical

Colors for binary variables: a vector of three colors representing TRUE, FALSE, and NA respectively (character).

colours_missing

Color for missing (NA) values in categorical plots (string).

show_legend_titles

Display titles for legends (flag).

legend_title_position

Position of the legend title ("top", "bottom", "left", "right").

legend_nrow

Number of rows in the legend (number).

legend_ncol

Number of columns in the legend. If set, legend_nrow should be NULL (number).

legend_title_size

Size of the legend title text (number).

legend_text_size

Size of the text within the legend (number).

legend_key_size

Size of the legend key symbols (number).

legend_orientation_heatmap

should legend orientation be "horizontal" or "vertical".

show_legend

Display the legend on the plot (flag).

legend_position

Position of the legend ("right", "left", "bottom", "top").

na_marker

Text used to mark NA values in numeric plots (string).

na_marker_size

Size of the text marker for NA values (number).

na_marker_colour

Color of the NA text marker (string).

show_na_marker_categorical

Show a marker for NA values on categorical tiles (flag).

show_na_marker_heatmap

Show a marker for NA values on heatmap tiles (flag).

colours_heatmap_low

Color for the lowest value in heatmaps (string).

colours_heatmap_high

Color for the highest value in heatmaps (string).

transform_heatmap

Transformation to apply before visualizing heatmap values ("identity", "log10", "log2").

fontsize_values_heatmap

Font size for heatmap values (number).

show_values_heatmap

Display numerical values on heatmap tiles (flag).

colours_values_heatmap

Color for heatmap values (string).

vertical_spacing

Space between each data row in points (number).

numeric_plot_type

Type of visualization for numeric data: "bar" or "heatmap".

y_axis_position

Position of the y-axis ("left" or "right").

width

controls how much space is present between bars and tiles within each plot. Can be 0-1 where values of 1 makes bars/tiles take up 100% of available space (no gaps between bars).

relative_height_numeric

how many times taller should numeric plots be relative to categorical tile plots. Only taken into account if numeric_plot_type == "bar" (number)

cli_header

Text used for h1 header. Included so it can be tweaked by packages that use ggstack, so they can customise how the info messages appear.

interactive_svg_width, interactive_svg_height

width and height of the interactive graphic region (in inches). Only used when interactive = TRUE.

fontsize_barplot_y_numbers

fontsize of the text describing numeric barplot max & min values (number).

max_digits_barplot_y_numbers

Number of digits to round the numeric barplot max and min values to (number).

fontsize_y_title

fontsize of the y axis titles (a.k.a the data.frame column names) (number).

beautify_text

Beautify y-axis text and legend titles by capitalizing words and adding spaces (flag).

Value

A list of visualization parameters for ggstack.

Examples


# Create Basic Plot
ggstack(baseballfans, col_id = "ID", col_sort = "Glasses")

# Configure plot ggstack_options()
ggstack(
  lazy_birdwatcher,
  col_sort = "Magpies",
  palettes = list(
    Birdwatcher = c(Robert = "#E69F00", Catherine = "#999999"),
    Day = c(Weekday = "#999999", Weekend = "#009E73")
  ),
  options = ggstack_options(
    show_legend = TRUE,
    fontsize_barplot_y_numbers = 12,
    legend_text_size = 16,
    legend_key_size = 1,
    legend_nrow = 1,
  )
)

Determine Whether Two Edges Cross

Description

Given the positions of two edges on the left and right axes, decides if they intersect in a parallel coordinates setup.

Usage

is_crossing(l1, r1, l2, r2)

Arguments

l1

Numeric position of the first edge on the left axis.

r1

Numeric position of the first edge on the right axis.

l2

Numeric position of the second edge on the left axis.

r2

Numeric position of the second edge on the right axis.

Value

A logical value. TRUE if they cross, FALSE otherwise.

Lazy Birdwatcher Dataset

Description

A simulated dataset describing the number of magpies observed by two birdwatchers.

Usage

lazy_birdwatcher

Format

lazy_birdwatcher

A data frame with 45 rows and 3 columns:

Magpies

Number of magpies observed

Day

Was the day of observation a weekday or a weekend?

Birdwatcher

Name of the birdwatcher

Dry Beans Dataset

Description

A subsample of the Koklu & Ozkan (2020) dry beans dataset produced by imaging a total of 13,611 grains from 7 varieties of dry beans. The original dataset contains 13,611 observations, but here we include a random subsample of 1000.

Usage

minibeans

Format

minibeans

A data frame with 1000 rows and 17 columns:

Area

The area of a bean zone and the number of pixels within its boundaries.

Perimeter

Bean circumference is defined as the length of its border.

Major axis length

The distance between the ends of the longest line that can be drawn from a bean.

Minor axis length

The longest line that can be drawn from the bean while standing perpendicular to the main axis.

Aspect ratio

Defines the relationship between L and l.

Eccentricity

Eccentricity of the ellipse having the same moments as the region.

Convex area

Number of pixels in the smallest convex polygon that can contain the area of a bean seed.

Equivalent diameter

The diameter of a circle having the same area as a bean seed area.

Extent

The ratio of the pixels in the bounding box to the bean area.

Solidity

Also known as convexity. The ratio of the pixels in the convex shell to those found in beans.

Roundness

Calculated with the following formula: (4piA)/(P^2).

Compactness

Measures the roundness of an object: Ed/L.

ShapeFactor1

Shape factor 1.

ShapeFactor2

Shape factor 2.

ShapeFactor3

Shape factor 3.

ShapeFactor4

Shape factor 4.

Class

Seker, Barbunya, Bombay, Cali, Dermosan, Horoz, and Sira.

Source

Koklu, M, and IA Ozkan. 2020. Multiclass Classification of Dry Beans Using Computer Vision and Machine Learning Techniques. Computers and Electronics in Agriculture, 174: 105507. doi: 10.1016/j.compag.2020.105507, https://doi.org/10.24432/C50S4B

Compute Mutual Information

Description

Computes mutual information between each feature in the features data frame and the target vector. The features are discretized using the "equalfreq" method from infotheo::discretize().

Usage

mutinfo(features, target, return_colnames = FALSE)

Arguments

features

A data frame of features. These will be discretized using the "equalfreq" method (see infotheo::discretize()).

target

A vector (character or factor) representing the variable to compute mutual information with.

return_colnames

Logical; if TRUE, returns the column names from features ordered by their mutual information with target (highest to lowest). If FALSE, returns mutual information values. (default: FALSE)

Value

If return_colnames = FALSE, a named numeric vector of mutual information scores is returned (one for each column in features), sorted in descending order. The names of the vector correspond to the column names of features. If return_colnames = TRUE, only the ordered column names of features are returned.

Examples

data(iris)
# Compute mutual information scores
mutinfo(iris[1:4], iris[[5]])

# Get column names ordered by mutual information with target column (most mutual info first)
mutinfo(iris[1:4], iris[[5]], return_colnames = TRUE)

Optimise the Ordering of Axes Using Distance Matrix

Description

Finds an ordering of axes that minimises a pairwise distance metric (usually the number of crossings). Offers brute-force and heuristic approaches.

Usage

optimise_axis_ordering_from_matrix(
  mx,
  method = c("auto", "brute_force", "repetitive_nn_with_2opt"),
  return_detailed = FALSE,
  verbose = TRUE
)

Arguments

mx

A matrix or dist object describing pairwise distances between axes.

method

A character string specifying the method. Can be "auto", "brute_force", or "repetitive_nn_with_2opt".

return_detailed

Logical; if TRUE, returns a list with detailed results for debugging.

verbose

Logical; if TRUE, prints progress messages.

Value

If return_detailed = FALSE, returns a character vector of axis names in the chosen order. Otherwise, returns a list with additional data.

Generate Permutations of the Integers 1..n

Description

Creates a matrix of all permutations for the integers from 1 to n.

Usage

permutations(n)

Arguments

n

Number of elements to permute.

Value

A matrix where each row is a permutation of 1..n.

Generate All Permutations of Axis Names

Description

Takes a character vector of axis names and returns a matrix of permutations.

Usage

permute_axis_names(axis_names)

Arguments

axis_names

A character vector of axis names.

Value

A matrix where each row represents one permutation of axis_names.

GGplot breaks

Description

Find sensible values to add 2 breaks at for a ggplot2 axis

Usage

sensible_2_breaks(vector)

Arguments

vector

vector fed into ggplot axis you want to define sensible breaks for

Value

vector of length 2. first element descripts upper break position, lower describes lower break