Skip to contents

Plot dataset comparison diagnostics

Source: R/PlotDataSetComparison.R
PlotDatasetComparison.Rd

Create diagnostic plots from a CompareDatasets() result object. This function visualizes dataset-version changes, including check status, summary metrics, structure changes, and variable-level value changes.

Usage

PlotDatasetComparison(
  CompareObj,
  Plot = c("All", "Checks", "SummaryMetrics", "StructureChanges", "VariableChanges",
    "TopChangedVariables"),
  interactive = TRUE,
  TopN = 10,
  Interactive = lifecycle::deprecated()
)

Arguments

CompareObj

A list returned by CompareDatasets().

Plot

Character value specifying which plot to return. Options are "All", "Checks", "SummaryMetrics", "StructureChanges", "VariableChanges", and "TopChangedVariables". Default is "All".

interactive

Logical; if TRUE, plots are converted to interactive plotly objects using plotly::ggplotly(). Default is TRUE.

TopN

Integer number of variables or records to preview in plots and hover text. Default is 10.

Interactive

Deprecated (since 19.15.0). Use interactive instead.

Value

If Plot = "All", a named list of plots. Otherwise, a single plot object. Plot objects are either ggplot objects or plotly htmlwidgets, depending on Interactive.

Details

Use this function after running CompareDatasets() to quickly inspect what changed between two versions of a dataset.

Examples

data(SampleData)

# Build two versions of a keyed dataset that differ in records, variables,
# and values, so every diagnostic panel has something to show.
old_data <- cbind(id = seq_len(nrow(SampleData)), SampleData)

new_data <- old_data[-(1:8), ]
new_data <- rbind(
  new_data,
  transform(old_data[1:3, ], id = max(old_data$id) + 1:3)
)
new_data$Cohort <- "Wave2"
new_data$Genotype <- NULL
new_data$age[1:25] <- new_data$age[1:25] + 1
new_data$Cortisol[1:15] <- new_data$Cortisol[1:15] * 1.2

comparison <- CompareDatasets(old_data, new_data, keys = "id")

diagnostics <- PlotDatasetComparison(
  comparison,
  Plot = "All",
  interactive = FALSE
)

# Check status, summary metrics, structural, value-change, and top-change views
diagnostics$Checks

diagnostics$SummaryMetrics

diagnostics$StructureChanges

diagnostics$VariableChanges

diagnostics$TopChangedVariables