Prepare a Check Result for Plotting

Coerces a check object — such as the result of check_na_timing() (from the anicheck package) — into the plot-ready data frame that its plot() method draws from. This is the staging step that sits between the diagnostic computation (the check_*() function) and the ggplot: it reshapes the natural, analysis-friendly columns of the check into the exact aesthetic contract a geom needs (a single grouping factor, ordered fill levels, the right level ordering, and so on).

Usage

as_plot_data(x, ...)

# Default S3 method
as_plot_data(x, ...)

# S3 method for class 'check_confidence'
as_plot_data(x, ..., clip = 0.02)

# S3 method for class 'check_na_gapsize'
as_plot_data(
  x,
  ...,
  ranked_by = c("occurrence", "total"),
  limit = 10,
  include_total = TRUE
)

# S3 method for class 'check_na_timing'
as_plot_data(x, ..., measure = c("percent", "count"), n_intervals = NULL)

Arguments

x

A check object.

...

Additional arguments passed to methods.

clip

For as_plot_data.check_confidence(): density floor (fraction of each keypoint's peak) below which the violin is cut. Default 0.02.

ranked_by, limit, include_total

For as_plot_data.check_na_gapsize(): order bars by "occurrence" or "total", keep the top limit per group, and whether to include the total-NAs series.

measure

For as_plot_data.check_na_timing(): "percent" (default, share of each interval) or "count" (number of frames).

n_intervals

For as_plot_data.check_na_timing(): number of time intervals to bin into (default NULL uses Sturges' rule on the largest group).

Value

A data frame classed for the corresponding plot method (for example anivis_check_na_timing_data).

Details

It is the anivis analog of data_plot() in the see package. Most users never call it directly — plot() calls it for you. Reach for it when you want the precise rows a plot is built from, to inspect them or to assemble a custom chart by hand.

Methods exist for check objects with a single canonical plot (one object class maps to one figure). Object types that can be plotted several different ways — an aniframe, which feeds both plot_trajectory() and plot_timeseries() — deliberately do not have a single method, since there would be nothing for it to dispatch on to choose between those shapes.

as_plot_data.check_confidence() turns the per-keypoint density grid into closed violin polygons: each keypoint sits at an integer y position, and its density is mirrored either side along the confidence (x) axis (width-normalised so every violin has the same maximum width). Grid points below clip x the peak are dropped, so thin tails and bimodal-bridging necks disappear and a split distribution becomes separate polygons. keypoint is the x-axis category; any other identity that varies (e.g. individual) collapses into a group factor for faceting. The x positions, a per-keypoint median / quartile overlay, and facet ride along as attributes. Returns a frame classed anivis_check_confidence_data.

as_plot_data.check_na_gapsize() reshapes the gap-size table into the long, ranked form the imputeTS-style bar chart needs: one row per (group, gap size, series), where series is occurrence (and total when include_total), value the count, and key a reorder_within-style factor ("<n> NA-gap___ <group>") ordered by ranked_by so each facet sorts independently. Returns a frame classed anivis_check_na_gapsize_data.

as_plot_data.check_na_timing() reconstructs, from the compact gap table, the missing / present frame counts per time interval — one row per (group, interval, status). A single interval width (in frames) is chosen for all groups (Sturges' rule by default) so bars line up across panels, and each gap's overlap with each interval is counted, so no per-frame data is needed. value is the share (measure = "percent") or count ("count"), width the interval's span in time units, and the frame interval size rides along as an attribute. Returns a frame classed anivis_check_na_timing_data.

See also

plot.anivis_check_na_timing()