politenessPlot {politeness}

Politeness plot

Description

Plots the prevalence of politeness features in documents, divided by a binary covariate.

Usage

politenessPlot(
  df_polite,
  split = NULL,
  split_levels = NULL,
  split_name = NULL,
  split_cols = c("firebrick", "navy"),
  top_title = "",
  drop_blank = 0.05,
  middle_out = 0.5,
  features = NULL,
  ordered = FALSE,
  CI = 0.68
)

Arguments

df_polite

a data.frame with politeness features calculated from a document set, as output by politeness.

split

a vector of covariate values. must have a length equal to the number of documents included in df_polite. No NA values allowed.

split_levels

character vector of length 2 default NULL. Labels for covariate levels for legend. If NULL, this will be inferred from split.

split_name

character default NULL. Name of the covariate for legend.

split_cols

character vector of length 2. Name of colors to use.

top_title

character default "". Title of plot.

drop_blank

Features less prevalent than this in the sample value are excluded from the plot. To include all features, set to 0

middle_out

Features less distinctive than this value (measured by p-value of t-test) are excluded. Defaults to 1 (i.e. include all).

features

character vector of feature names. If NULL all will be included.

ordered

logical should features be ordered according to features param? default is FALSE.

CI

Coverage of error bars. Defaults to 0.68 (i.e. standard error).

Details

Length of split must be the same as number of rows of df_polite. Typically split should be a two-category variable. However, if a continuous covariate is given, then the top and bottom terciles of that distribution are treated as the two categories (while dropping data from the middle tercile).

Value

a ggplot of the prevalence of politeness features, conditional on split. Features are sorted by variance-weighted log odds ratio.

Examples

data("phone_offers")

polite.data<-politeness(phone_offers$message, parser="none", drop_blank=FALSE)

politeness::politenessPlot(polite.data,
                           split=phone_offers$condition,
                           split_levels = c("Tough","Warm"),
                           split_name = "Condition",
                           top_title = "Average Feature Counts")


politeness::politenessPlot(polite.data,
                           split=phone_offers$condition,
                           split_levels = c("Tough","Warm"),
                           split_name = "Condition",
                           top_title = "Average Feature Counts",
                           features=c("Positive.Emotion","Hedges","Negation"))


polite.data<-politeness(phone_offers$message, parser="none", metric="binary", drop_blank=FALSE)

politeness::politenessPlot(polite.data,
                           split=phone_offers$condition,
                           split_levels = c("Tough","Warm"),
                           split_name = "Condition",
                           top_title = "Binary Feature Use")