Package 'volker'

ChatGPT Adoption Dataset CG-GE-APR23

Description

A small random subset of data from a survey about ChatGPT adoption. The survey was conducted in April 2023 within the population of German Internet users.

Usage

chatgpt

Format

chatgpt

A data frame with 101 rows and 19 columns:

case

A running case number

adopter

Adoption groups inspired by Roger's innovator typology.

use_

Columns starting with use contain data about ChatGPT usage in different contexts.

cg_activities

Text answers to the question, what the respondents do with ChatGPT.

cg_adoption_

A scale consisting of items about advantages, fears, and social aspects. The scales match theoretical constructs inspired by Roger's diffusion model and Davis' Technology Acceptance Model

sd_

Columns starting with sd contain sociodemographics of the respondents.

Details

Call codebook(volker::chatgpt) to see the items and and answer options.

Source

Communication Department of the University of Münster ([email protected]).

---

Get variable labels from their comment attributes

Description

Usage

codebook(data, cols)

Arguments

data	A tibble.
cols	A tidy variable selections to filter specific columns.

Value

A tibble with the columns: - item_name: The column name. - item_group: First part of the column name, up to an underscore. - item_class: The last class value of an item (e.g. numeric, factor). - item_label: The comment attribute of the column. - value_name: In case a column has numeric attributes, the attribute names. - value_label: In case a column has numeric attributes or T/F-attributes, the attribute values. In case a column has a levels attribute, the levels.

Examples

volker::codebook(volker::chatgpt)

---

Output effect sizes and test statistics for count data

Description

The type of effect size depends on the number of selected columns:

• One categorical column: see effect_counts_one

• Multiple categorical columns: see effect_counts_items

Cross tabulations:

• One categorical column and one grouping column: see effect_counts_one_grouped

• Multiple categorical columns and one grouping column: see effect_counts_items_grouped (not yet implemented)

• Multiple categorical columns and multiple grouping columns: effect_counts_items_grouped_items (not yet implemented)

By default, if you provide two column selections, the second column is treated as categorical. Setting the metric-parameter to TRUE will call the appropriate functions for correlation analysis:

• One categorical column and one metric column: see effect_counts_one_cor (not yet implemented)

• Multiple categorical columns and one metric column: see effect_counts_items_cor (not yet implemented)

• Multiple categorical columns and multiple metric columns:effect_counts_items_cor_items (not yet implemented)

Usage

effect_counts(data, cols, cross = NULL, metric = FALSE, clean = TRUE, ...)

Arguments

data	A data frame.
cols	A tidy column selection, e.g. a single column (without quotes) or multiple columns selected by methods such as starts_with().
cross	Optional, a grouping column. The column name without quotes.
metric	When crossing variables, the cross column parameter can contain categorical or metric values. By default, the cross column selection is treated as categorical data. Set metric to TRUE, to treat it as metric and calculate correlations.
clean	Prepare data by data_clean.
...	Other parameters passed to the appropriate effect function.

Value

A volker tibble.

Examples

library(volker)
data <- volker::chatgpt

effect_counts(data, sd_gender, adopter)

---

Output effect sizes and test statistics for metric data

Description

The calculations depend on the number of selected columns:

• One metric column: see effect_metrics_one

• Multiple metric columns: see effect_metrics_items

Group comparisons:

• One metric column and one grouping column: see effect_metrics_one_grouped

• Multiple metric columns and one grouping column: see effect_metrics_items_grouped

• Multiple metric columns and multiple grouping columns: not yet implemented

By default, if you provide two column selections, the second column is treated as categorical. Setting the metric-parameter to TRUE will call the appropriate functions for correlation analysis:

• Two metric columns: see effect_metrics_one_cor

• Multiple metric columns and one metric column: see effect_metrics_items_cor

• Two metric column selections: see effect_metrics_items_cor_items

Usage

effect_metrics(data, cols, cross = NULL, metric = FALSE, clean = TRUE, ...)

Arguments

data	A data frame.
cols	A tidy column selection, e.g. a single column (without quotes) or multiple columns selected by methods such as starts_with().
cross	Optional, a grouping column (without quotes).
metric	When crossing variables, the cross column parameter can contain categorical or metric values. By default, the cross column selection is treated as categorical data. Set metric to TRUE, to treat it as metric and calculate correlations.
clean	Prepare data by data_clean.
...	Other parameters passed to the appropriate effect function.

Value

A volker tibble.

Examples

library(volker)
data <- volker::chatgpt

effect_metrics(data, sd_age, sd_gender)

---

Volker style HTML document format

Description

Based on the standard theme, tweaks the pill navigation to switch between tables and plots. To use the format, in the header of your Markdown document, set output: volker::html_report.

Usage

html_report(...)

Arguments

...	Additional arguments passed to html_document.

Value

R Markdown output format.

Examples

## Not run: 
# Add `volker::html_report` to the output options of your Markdown document:
#
# ```
# ---
# title: "How to create reports?"
# output: volker::html_report
# ---
# ```

## End(Not run)

---

Calculate the mean value of multiple items

Description

Usage

idx_add(data, cols, newcol = NULL, reverse = NULL, clean = TRUE)

Arguments

data	A dataframe.
cols	A tidy selection of item columns.
newcol	Name of the index as a character value. Set to NULL (default) to automatically build a name from the common column prefix, prefixed with "idx_".
reverse	A tidy selection of columns with reversed codings. For example, if you want to calculate an index of the two items "I feel bad about this" and "I like it", both coded with 1=not at all to 5=fully agree, you need to reverse one of them to make the codings compatible.
clean	Prepare data by data_clean.

Value

The input tibble with an additional column that contains the index values. The column contains the result of the alpha calculation in the attribute named "psych.alpha".

Examples

ds <- volker::chatgpt
volker::idx_add(ds, starts_with("cg_adoption"))

---

Set column and value labels

Description

Usage

labs_apply(data, codes = NULL, cols = NULL, items = TRUE, values = TRUE)

Arguments

data	A tibble containing the dataset.
codes	A tibble in codebook format.
cols	A tidy column selection. Set to NULL (default) to apply to all columns found in the codebook. Restricting the columns is helpful when you want to set value labels. In this case, provide a tibble with value_name and value_label columns and specify the columns that should be modified.
items	If TRUE, column labels will be retrieved from the codes (the default). If FALSE, no column labels will be changed. Alternatively, a named list of column names with their labels.
values	If TRUE, value labels will be retrieved from the codes (default). If FALSE, no value labels will be changed. Alternatively, a named list of value names with their labels. In this case, use the cols-Parameter to define which columns should be changed.

Details

You can either provide a data frame in codebook format to the codes-parameter or provide named lists to the items- or values-parameter.

When working with a codebook in the codes-parameter:

• Change column labels by providing the columns item_name and item_label in the codebook. Set the items-parameter to TRUE (the default setting).

• Change value labels by providing the columns value_name and value_label in the codebook. To tell which columns should be changed, you can either use the item_name column in the codebook or use the cols-parameter. For factor values, the levels and their order are retrieved from the value_label column. For coded values, labels are retrieved from both the columns value_name and value_label.

When working with lists in the items- or values-parameter:

• Change column labels by providing a named list to the items-parameter. The list contains labels named by the columns. Set the parameters codes and cols to NULL (their default value).

• Change value labels by providing a named list to the values-parameter. The list contains labels named by the values. Provide the column selection in the cols-parameter. Set the codes-parameter to NULL (its default value).

Value

A tibble containing the dataset with new labels.

Examples

library(volker)

# Set column labels using the items-parameter
volker::chatgpt %>%
  labs_apply(
   items = list(
     "cg_adoption_advantage_01" = "Allgemeine Vorteile",
     "cg_adoption_advantage_02" = "Finanzielle Vorteile",
     "cg_adoption_advantage_03" = "Vorteile bei der Arbeit",
     "cg_adoption_advantage_04" = "Macht mehr Spaß"
   )
 ) %>%
 tab_metrics(starts_with("cg_adoption_advantage_"))

# Set value labels using the values-parameter
 volker::chatgpt %>%
   labs_apply(
     cols=starts_with("cg_adoption"),
     values = list(
       "1" = "Stimme überhaupt nicht zu",
       "2" = "Stimme nicht zu",
       "3" = "Unentschieden",
       "4" = "Stimme zu",
       "5" =  "Stimme voll und ganz zu"
     )
   ) %>%
   plot_metrics(starts_with("cg_adoption"))

---

Remove all comments from the selected columns

Description

Usage

labs_clear(data, cols, labels = NULL)

Arguments

data	A tibble.
cols	Tidyselect columns.
labels	The attributes to remove. NULL to remove all attributes except levels and class.

Value

A tibble with comments removed.

Examples

library(volker)
volker::chatgpt |>
  labs_clear()

---

Restore labels from the codebook store in the codebook attribute.

Description

Usage

labs_restore(data, cols = NULL)

Arguments

data	A data frame.
cols	A tidyselect column selection.

Details

You can store labels before mutate operations by calling labs_store.

Value

A data frame.

Examples

library(dplyr)
library(volker)

volker::chatgpt |>
  labs_store() |>
  mutate(sd_age = 2024 - sd_age) |>
  labs_restore() |>
  tab_metrics(sd_age)

---

Get the current codebook and store it in the codebook attribute.

Description

Usage

labs_store(data)

Arguments

data	A data frame.

Details

You can restore the labels after mutate operations by calling labs_restore.

Value

A data frame.

Examples

library(dplyr)
library(volker)

volker::chatgpt |>
  labs_store() |>
  mutate(sd_age = 2024 - sd_age) |>
  labs_restore() |>
  tab_metrics(sd_age)

---

Volker style PDF document format

Description

Based on the standard theme, tweaks tex headers. To use the format, in the header of your Markdown document, set output: volker::pdf_report.

Usage

pdf_report(...)

Arguments

...	Additional arguments passed to pdf_document.

Value

R Markdown output format.

Examples

## Not run: 
# Add `volker::pdf_report` to the output options of your Markdown document:
#
# ```
# ---
# title: "How to create reports?"
# output: volker::pdf_report
# ---
# ```

## End(Not run)

---

Output a frequency plot

Description

The type of frequency plot depends on the number of selected columns:

• One categorical column: see plot_counts_one

• Multiple categorical columns: see plot_counts_items

Cross tabulations:

• One categorical column and one grouping column: see plot_counts_one_grouped

• Multiple categorical columns and one grouping column: see plot_counts_items_grouped

• Two categorical column selections: see plot_counts_items_grouped_items (not yet implemented)

By default, if you provide two column selections, the second selection is treated as categorical. Setting the metric-parameter to TRUE will call the appropriate functions for correlation analysis:

• One categorical column and one metric column: see plot_counts_one_cor

• Multiple categorical columns and one metric column: see plot_counts_items_cor

• Multiple categorical columns and multiple metric columns: see plot_counts_items_cor_items (not yet implemented)

Parameters that may be passed to the count functions (see the respective function help):

• ci: Add confidence intervals to proportions.

• ordered: The values of the cross column can be nominal (0), ordered ascending (1), or ordered descending (-1). The colors are adjusted accordingly.

• category: When you have multiple categories in a column, you can focus one of the categories to simplify the plots. By default, if a column has only TRUE and FALSE values, the outputs focus the TRUE category.

• prop: For stacked bar charts, displaying row percentages instead of total percentages gives a direct visual comparison of groups.

• limits: The scale limits are automatically guessed by the package functions (work in progress). Use the limits-parameter to manually fix any misleading graphs.

• title: All plots usually get a title derived from the column attributes or column names. Set to FALSE to suppress the title or provide a title of your choice as a character value.

• labels: Labels are extracted from the column attributes. Set to FALSE to output bare column names and values.

• numbers: Set the numbers parameter to “n” (frequency), “p” (percentage) or c(“n”,“p”). To prevent cluttering and overlaps, numbers are only plotted on bars larger than 5%.

Usage

plot_counts(data, cols, cross = NULL, metric = FALSE, clean = TRUE, ...)

Arguments

data	A data frame.
cols	A tidy column selection, e.g. a single column (without quotes) or multiple columns selected by methods such as starts_with().
cross	Optional, a grouping column. The column name without quotes.
metric	When crossing variables, the cross column parameter can contain categorical or metric values. By default, the cross column selection is treated as categorical data. Set metric to TRUE, to treat it as metric and calculate correlations.
clean	Prepare data by data_clean.
...	Other parameters passed to the appropriate plot function.

Value

A ggplot2 plot object.

Examples

library(volker)
data <- volker::chatgpt

plot_counts(data, sd_gender)

---

Output a plot with distribution parameters such as the mean values

Description

The plot type depends on the number of selected columns:

• One metric column: see plot_metrics_one

• Multiple metric columns: see plot_metrics_items

Group comparisons:

• One metric column and one grouping column: see plot_metrics_one_grouped

• Multiple metric columns and one grouping column: see plot_metrics_items_grouped

• Multiple metric columns and multiple grouping columns: see plot_metrics_items_grouped_items (not yet implemented)

By default, if you provide two column selections, the second selection is treated as categorical. Setting the metric-parameter to TRUE will call the appropriate functions for correlation analysis:

• Two metric columns: see plot_metrics_one_cor

• Multiple metric columns and one metric column : see plot_metrics_items_cor

• Two metric column selections: see plot_metrics_items_cor_items

Parameters that may be passed to the metric functions (see the respective function help):

• ci: Plot confidence intervals for means or correlation coefficients.

• box: Visualise the distribution by adding boxplots.

• log: In scatter plots, you can use a logarithmic scale. Be aware, that zero values will be omitted because their log value is undefined.

• method: By default, correlations are calculated using Pearson’s R. You can choose Spearman’s Rho with the methods-parameter.

• limits: The scale limits are automatically guessed by the package functions (work in progress). Use the limits-parameter to manually fix any misleading graphs.

• title: All plots usually get a title derived from the column attributes or column names. Set to FALSE to suppress the title or provide a title of your choice as a character value.

• labels: Labels are extracted from the column attributes. Set to FALSE to output bare column names and values.

• numbers: Controls whether to display correlation coefficients on the plot.

Usage

plot_metrics(data, cols, cross = NULL, metric = FALSE, clean = TRUE, ...)

Arguments

data	A data frame.
cols	A tidy column selection, e.g. a single column (without quotes) or multiple columns selected by methods such as starts_with().
cross	Optional, a grouping column (without quotes).
metric	When crossing variables, the cross column parameter can contain categorical or metric values. By default, the cross column selection is treated as categorical data. Set metric to TRUE, to treat it as metric and calculate correlations.
clean	Prepare data by data_clean.
...	Other parameters passed to the appropriate plot function.

Value

A ggplot object.

Examples

library(volker)
data <- volker::chatgpt

plot_metrics(data, sd_age)

---

Create table and plot for categorical variables

Description

Depending on your column selection, different types of plots and tables are generated. See plot_counts and tab_counts.

Usage

report_counts(
  data,
  cols,
  cross = NULL,
  metric = FALSE,
  index = FALSE,
  effect = FALSE,
  numbers = NULL,
  title = TRUE,
  close = TRUE,
  clean = TRUE,
  ...
)

Arguments

data	A data frame.
cols	A tidy column selection, e.g. a single column (without quotes) or multiple columns selected by methods such as starts_with().
cross	Optional, a grouping column (without quotes).
metric	When crossing variables, the cross column parameter can contain categorical or metric values. By default, the cross column selection is treated as categorical data. Set metric to TRUE, to treat it as metric and calculate correlations.
index	When the cols contain items on a metric scale (as determined by get_direction), an index will be calculated using the 'psych' package. Set to FALSE to suppress index generation.
effect	Whether to report statistical tests and effect sizes. See effect_counts for further parameters.
numbers	The numbers to print on the bars: "n" (frequency), "p" (percentage) or both. Set to NULL to remove numbers.
title	A character providing the heading or TRUE (default) to output a heading. Classes for tabset pills will be added.
close	Whether to close the last tab (default value TRUE) or to keep it open. Keep it open to add further custom tabs by adding headers on the fifth level in Markdown (e.g. ##### Method).
clean	Prepare data by data_clean.
...	Parameters passed to the plot_counts and tab_counts and effect_counts functions.

Details

For item batteries, an index is calculated and reported. When used in combination with the Markdown-template "html_report", the different parts of the report are grouped under a tabsheet selector.

Value

A volker report object.

Examples

library(volker)
data <- volker::chatgpt

report_counts(data, sd_gender)

---

Create table and plot for metric variables

Description

Depending on your column selection, different types of plots and tables are generated. See plot_metrics and tab_metrics.

Usage

report_metrics(
  data,
  cols,
  cross = NULL,
  metric = FALSE,
  ...,
  index = FALSE,
  effect = FALSE,
  title = TRUE,
  close = TRUE,
  clean = TRUE
)

Arguments

data	A data frame.
cols	A tidy column selection, e.g. a single column (without quotes) or multiple columns selected by methods such as starts_with().
cross	Optional, a grouping or correlation column (without quotes).
metric	When crossing variables, the cross column parameter can contain categorical or metric values. By default, the cross column selection is treated as categorical data. Set metric to TRUE, to treat it as metric and calculate correlations.
...	Parameters passed to the plot_metrics and tab_metrics and effect_metrics functions.
index	When the cols contain items on a metric scale (as determined by get_direction), an index will be calculated using the 'psych' package. Set to FALSE to suppress index generation.
effect	Whether to report statistical tests and effect sizes. See effect_counts for further parameters.
title	A character providing the heading or TRUE (default) to output a heading. Classes for tabset pills will be added.
close	Whether to close the last tab (default value TRUE) or to keep it open. Keep it open to add further custom tabs by adding headers on the fifth level in Markdown (e.g. ##### Method).
clean	Prepare data by data_clean.

Details

For item batteries, an index is calculated and reported. When used in combination with the Markdown-template "html_report", the different parts of the report are grouped under a tabsheet selector.

Value

A volker report object.

Examples

library(volker)
data <- volker::chatgpt

report_metrics(data, sd_age)

---

Output a frequency table

Description

The type of frequency table depends on the number of selected columns:

• One categorical column: see tab_counts_one

• Multiple categorical columns: see tab_counts_items

Cross tabulations:

• One categorical column and one grouping column: see tab_counts_one_grouped

• Multiple categorical columns and one grouping column: see tab_counts_items_grouped

• Multiple categorical columns and multiple grouping columns: see tab_counts_items_grouped_items (not yet implemented)

By default, if you provide two column selections, the second column is treated as categorical. Setting the metric-parameter to TRUE will call the appropriate functions for correlation analysis:

• One categorical column and one metric column: see tab_counts_one_cor

• Multiple categorical columns and one metric column: see tab_counts_items_cor

• Multiple categorical columns and multiple metric columns: tab_counts_items_cor_items (not yet implemented)

Parameters that may be passed to specific count functions:

• ci: Add confidence intervals to proportions.

• percent: Frequency tables show percentages by default. Set to FALSE to get raw proportions.

• prop: For cross tables you can choose between total, row or column percentages.

• values: The values to output: n (frequency) or p (percentage) or both (the default).

• category: When you have multiple categories in a column, you can focus one of the categories to simplify the plots. By default, if a column has only TRUE and FALSE values, the outputs focus the TRUE category.

• labels: Labels are extracted from the column attributes. Set to FALSE to output bare column names and values.

Usage

tab_counts(data, cols, cross = NULL, metric = FALSE, clean = TRUE, ...)

Arguments

data	A data frame.
cols	A tidy column selection, e.g. a single column (without quotes) or multiple columns selected by methods such as starts_with().
cross	Optional, a grouping column. The column name without quotes.
metric	When crossing variables, the cross column parameter can contain categorical or metric values. By default, the cross column selection is treated as categorical data. Set metric to TRUE, to treat it as metric and calculate correlations.
clean	Prepare data by data_clean.
...	Other parameters passed to the appropriate table function.

Value

A volker tibble.

Examples

library(volker)
data <- volker::chatgpt

tab_counts(data, sd_gender)

---

Output a table with distribution parameters

Description

The table type depends on the number of selected columns:

• One metric column: see tab_metrics_one

• Multiple metric columns: see tab_metrics_items

Group comparisons:

• One metric column and one grouping column: see tab_metrics_one_grouped

• Multiple metric columns and one grouping column: see tab_metrics_items_grouped

• Multiple metric columns and multiple grouping columns: see tab_metrics_items_grouped_items (not yet implemented)

By default, if you provide two column selections, the second column is treated as categorical. Setting the metric-parameter to TRUE will call the appropriate functions for correlation analysis:

• Two metric columns: see tab_metrics_one_cor

• Multiple metric columns and one metric column: see tab_metrics_items_cor

• Two metric column selections: see tab_metrics_items_cor_items

Parameters that may be passed to specific metric functions:

• ci: Add confidence intervals for means or correlation coefficients.

• values: The output metrics, mean (m), the standard deviation (sd) or both (the default).

• digits: Tables containing means and standard deviations by default round values to one digit. Increase the number to show more digits

• method: By default, correlations are calculated using Pearson’s R. You can choose Spearman’s Rho with the methods-parameter.

• labels: Labels are extracted from the column attributes. Set to FALSE to output bare column names and values.

Usage

tab_metrics(data, cols, cross = NULL, metric = FALSE, clean = TRUE, ...)

Arguments

data	A data frame.
cols	A tidy column selection, e.g. a single column (without quotes) or multiple columns selected by methods such as starts_with().
cross	Optional, a grouping column (without quotes).
metric	When crossing variables, the cross column parameter can contain categorical or metric values. By default, the cross column selection is treated as categorical data. Set metric to TRUE, to treat it as metric and calculate correlations.
clean	Prepare data by data_clean.
...	Other parameters passed to the appropriate table function.

Value

A volker tibble.

Examples

library(volker)
data <- volker::chatgpt

tab_metrics(data, sd_age)

---

Define a default theme for volker plots

Description

Set ggplot colors, sizes and layout parameters.

Usage

theme_vlkr(
  base_size = 11,
  base_color = "black",
  base_fill = VLKR_FILLDISCRETE,
  base_gradient = VLKR_FILLGRADIENT
)

Arguments

base_size	Base font size.
base_color	Base font color.
base_fill	A list of fill color sets or at least one fill color set. Example: list(c("red"), c("red", "blue", "green")). Each set can contain different numbers of colors. Depending on the number of colors needed, the set with at least the number of required colors is used. The first color is always used for simple bar charts.
base_gradient	A color vector used for creating gradient fill colors, e.g. in stacked bar plots.

Details

Value

A theme function.

Examples

library(volker)
library(ggplot2)
data <- volker::chatgpt

theme_set(theme_vlkr(base_size=15, base_fill = list("red")))
plot_counts(data, sd_gender)

---