Create a one-way analysis plot — one

Bins a feature variable, computes exposure-weighted means of one or more observed variables per bin, and returns a dual-axis plotly chart (bars = exposure, lines = weighted means). Optionally splits by a second variable.

Usage

one_way(data, ...)

# Default S3 method
one_way(
  data,
  var,
  obs = "target",
  exposure = "exposure",
  split = NA_character_,
  bins = 35L,
  time_unit = NA_character_,
  type_agg = c("equal_exposure", "equal_range"),
  ret = c("plot", "data"),
  verbose = FALSE,
  ...
)

Arguments

data: A data.frame or data.table.
...: Arguments passed to methods.
var: Column to plot on the x-axis. Bare name or string.
obs: One or more columns to summarise on the y-axis (right axis). Bare name, c(a, b), or character vector. Default "target".
exposure: Column of exposure weights. If the column does not exist in data, every row is given weight 1. Bare name or string. Default "exposure".
split: Optional column to split lines by. Bare name, string, or NA / NULL for no split. Default NA.
bins: [integer(1)] Number of equal-exposure bins for numeric variables with more than bins unique values. Default 35.
time_unit: [character(1)] For Date / POSIXct variables, the width of each bin. Passed directly to base::cut.POSIXt(), so any string it accepts works: "month", "week", "2 weeks", "12 hours", "quarter", "year", etc. Ignored for non-date columns. Default NA (no date binning — dates are shown as individual values).
type_agg: [character(1)] Binning strategy for numeric variables: "equal_exposure" (default) or "equal_range".
ret: [character(1)] "plot" (default) returns a plotly object; "data" returns the aggregated data.table.
verbose: [logical(1)] Report how many NA values of var were moved to the trailing "NA" category? Default FALSE.

Value

A plotly object, or a data.table when ret = "data", or NULL with a warning if the plot cannot be produced.

Details

Column name arguments (var, obs, exposure, split) accept both bare (unquoted) names and strings, so one_way(df, wt, mpg) and one_way(df, "wt", "mpg") are equivalent.

Examples

# \donttest{
# bare names
one_way(mtcars, wt, mpg, bins = 10)

one_way(mtcars, cyl, c(mpg, hp), split = am)


# strings (equivalent)
one_way(mtcars, var = "wt", obs = "mpg", bins = 10)

one_way(mtcars, var = "cyl", obs = c("mpg", "hp"), split = "am")

# }