Missingness Map

Visualizes missingness across a data frame as a heatmap, either per row/column or aggregated by a grouping column.

Usage

missmap(
  df,
  by_column = NULL,
  na_list = NULL,
  row_order = FALSE,
  fill = c("auto", "binary", "count")
)

Arguments

df

A data frame or tibble.

by_column

Optional unquoted column name to aggregate missingness by (e.g. a site or subject id) instead of plotting every row individually. Missingness is summed per feature within each group. When row_order = FALSE, both the groups (x-axis) and the features (y-axis) are ordered by descending total missingness, exactly as in the per-row/column view.

na_list

Optional vector of additional values to treat as missing (beyond actual NA).

row_order

Logical; if FALSE (default) rows/features (or, in by_column mode, groups/features) are ordered by how much missingness they have.

fill

In by_column mode, how to color the tiles: "auto" (default) picks a discrete two-level "Missing"/"Present" fill (matching the per-row/column view) when every group has at most one contributing row – in which case the summed missingness count is always 0 or 1 and a count scale would be misleading – and otherwise falls back to a continuous "# missing" gradient. "binary" and "count" force one of those two behaviors regardless of group size. Ignored outside by_column mode.

Value

A ggplot object.

Examples

df <- data.frame(
  a = c(1, NA, 3, 4, NA),
  b = c(NA, NA, 3, 4, 5),
  site = c("A", "A", "B", "B", "B")
)
missmap(df)

missmap(df, by_column = site)