Package {vizintel}

Plot recommended charts from a viz_recommend_result

Description

Renders each recommended chart type as a real ggplot using the data stored inside the result object, then arranges all panels into a single view with gridExtra::grid.arrange(). Each panel is labelled with the chart type and its fit score.

Usage

## S3 method for class 'viz_recommend_result'
plot(x, ...)

Arguments

Details

Call plot() directly on the object returned by viz_recommend() – the data is stored automatically so you do not need to pass it again.

Value

Invisibly returns NULL. Called for its side-effect of rendering the arranged plot grid.

Examples

res <- viz_recommend(mtcars["mpg"])
plot(res)

# Paired mode -- scatter / hexbin arranged together
res2 <- viz_recommend(mtcars, x = "wt", y = "mpg")
plot(res2)

Print a viz_audit_report

Description

Prints each detected issue with its severity prefix and technical message, followed by a plain-English explanation (the explanation field) on the next line prefixed with "In plain terms: ".

Usage

## S3 method for class 'viz_audit_report'
print(x, ...)

Arguments

Value

x, invisibly.

Print a viz_recommend_result

Description

Print a viz_recommend_result

Usage

## S3 method for class 'viz_recommend_result'
print(x, ...)

Arguments

Value

x, invisibly.

Profile a data frame's columns

Description

An internal engine that inspects every column of a data frame, classifies it as "numeric", "categorical", or "datetime", and computes type-specific summary statistics. The result is a tidy data frame with one row per variable – intended to feed downstream visualization-recommendation logic.

Usage

profile_data(df)

Arguments

Value

A data.frame with one row per column of df and the following nine columns:

Summarise a viz_audit_report

Description

Summarise a viz_audit_report

Usage

## S3 method for class 'viz_audit_report'
summary(object, ...)

Arguments

Value

object, invisibly.

Check visualization accessibility

Description

Checks an existing ggplot for common visualization accessibility and readability issues, including excessive colours, colour-only encoding, excessive shapes, small text, excessive text labels, and overplotting.

Usage

viz_accessibility(plot)

Arguments

Value

A list containing a summary and an issues element. Each issue contains a severity level, check name, and descriptive message.

Examples

library(ggplot2)

p <- ggplot(mtcars, aes(x = wt, y = mpg, colour = factor(cyl))) +
  geom_point()

viz_accessibility(p)

Suggest alternative visualizations

Description

Suggests and generates alternative visualization types for an existing ggplot object. The generated alternatives are displayed via the active graphics device and are also returned for further use. Informational messages about the alternatives can be suppressed with suppressMessages().

Usage

viz_alternatives(plot)

Arguments

Value

A list containing:

recommendations

Character vector of suggested alternative visualization names.

plots

A named list of generated ggplot objects for the alternatives.

Examples

library(ggplot2)

p <- ggplot(mtcars, aes(wt, mpg)) +
  geom_point()

result <- suppressMessages(viz_alternatives(p))
result$recommendations

Audit a ggplot object for common visualization integrity problems

Description

Inspects a ggplot object by rendering it internally via ggplot2::ggplot_build() and running four integrity checks drawn from data-visualization best practices (Tufte, Cairo, Wilke). Returns a structured S3 report with one entry per detected issue.

Usage

viz_audit(plot)

Arguments

Details

Checks performed

Lie factor note

The lie factor is computed in data space (displayed axis range vs actual data range), not pixel space. Pixel-based computation requires a live graphics device and is not reproducible in a non-interactive context.

Plain-English explanations

Every issue includes an explanation field with a plain-English description of what the problem means for how misleading the chart looks. This is printed after the technical message when you call print() on the report.

Value

An S3 object of class "viz_audit_report", which is a list with:

n_issues

Total number of issues detected (integer).

n_critical

Number of critical issues (integer).

n_warning

Number of warning-level issues (integer).

n_info

Number of info-level issues (integer).

issues

A list of issue lists, each with fields check, severity, message, details, and explanation.

plot_geoms

Character vector of geom class names in the plot.

Examples

library(ggplot2)

# A bar chart with a truncated y-axis -- should flag "truncated_axis"
bad <- ggplot(data.frame(x = c("A", "B"), y = c(50, 80)), aes(x, y)) +
  geom_col() +
  coord_cartesian(ylim = c(40, 85))
report <- viz_audit(bad)
print(report)

# A clean scatter plot -- should return no critical issues
good <- ggplot(mtcars, aes(wt, mpg)) + geom_point()
viz_audit(good)

Explore Relationships Between Numeric Variables

Description

Explores relationships between numeric variables using correlation analysis and a heatmap.

Usage

viz_explore(data)

Arguments

Value

A list containing:

• correlation: A correlation matrix.

• plot: A ggplot2 correlation heatmap.

Examples

result <- viz_explore(mtcars)
result$correlation

result$plot

Recommend chart types for a data frame's columns

Description

Calls profile_data() internally to understand each column, then scores candidate chart types from an internal taxonomy against the column profile and any goal or variable hint the user provides. Returns a ranked list of recommendations, highest scoring first.

Usage

viz_recommend(df, goal = NULL, x = NULL, y = NULL)

Arguments

Details

How scoring works

An internal taxonomy table lists chart types alongside the conditions under which each chart is a good fit (variable types, goal, skewness, cardinality, data size). Each applicable chart starts with a base score and is adjusted:

• +0.15 if the goal matches exactly rather than being a generic match

• -0.20 if the data is skewed (|skewness| > 1) and the chart does not handle skew well

• -0.20 if the column has more than 15 unique categories and the chart becomes unreadable at that cardinality

• -0.15 if there are more than 1000 rows and the chart does not scale well

Charts with a final score below 0.10 are dropped. The top 3 per variable/pair are returned.

Taxonomy overview

Value

A list of class "viz_recommend_result". Each element is a list of class "viz_recommendation" with fields:

variable

Column name, or "x vs y" string for pairs.

chart

Chart type string.

fit_score

Numeric in [0, 1]. Higher is better.

rationale

One sentence explaining why this chart fits.

goal

The goal used when scoring this recommendation.

The result also has attributes df, x, and y (the data and variable names stored so that plot() can be called on the result directly, without passing data again).

Side effect: when called in an interactive session, the recommended charts are automatically rendered in the Plots pane via plot.viz_recommend_result(). You can also call plot() manually on the returned object at any time.

Examples

# Per-column recommendations for the whole data frame
viz_recommend(mtcars)


# Two specific columns as a pair
viz_recommend(mtcars, x = "wt", y = "mpg")

# Force a goal
viz_recommend(mtcars["mpg"], goal = "distribution")