| Title: | Helper functions for pretty-printing numbers |
|---|---|
| Description: | This package contains helper functions for formatting numbers. |
| Authors: | Tristan Mahr [aut, cre] |
| Maintainer: | Tristan Mahr <[email protected]> |
| License: | GPL-3 + file LICENSE |
| Version: | 0.0.0.9005 |
| Built: | 2024-10-27 04:19:32 UTC |
| Source: | https://github.com/tjmahr/printy |
Format an effect from a model object in markdown
fmt_effect_md( model, effect, terms = "besp", digits = 2, statistic = NULL, b_lab = NULL, ci_width = 0.95, p_value_method = NULL )fmt_effect_md( model, effect, terms = "besp", digits = 2, statistic = NULL, b_lab = NULL, ci_width = 0.95, p_value_method = NULL )
model |
a model object |
effect |
string naming an effect from a model |
terms |
a string representing the terms about the effect to extract and
format and the order to print the terms. See details below. Defaults to
|
digits |
a vector of digits to use for non-p-value terms. Defaults to 2 for 2 decimal places of precision for all terms. This argument can be a vector to set the digits for each term, but in this case, the digits is still ignored for p-values. |
statistic |
symbol to use for statistic. Defaults to t (or z in glmer models). |
b_lab |
label to print in subscripts after b for when |
ci_width |
width to use for confidence intervals when the term |
Currently only effects fit by stats::lm() and lme4::lmer().
The supported terms are:
"b" - parameter estimate (think b for beta)
"B" - parameter estimate with a subscript label provided by b_lab
"e" - standard error
"s" - statistic. The symbol for the statistic is set by
statistic. The default value is "t" for a t-statistic. Example
output: t = 1.
"S" - statistic as in "s" but with degrees of freedom. Example
output: t(12) = 1.
"i" - confidence interval. Width is set by ci_width.
"p" - p-value. The p-value is formatted by fmt_p_value_md().
Degrees of freedom and p-values for lmer() models use the
Kenwood-Rogers approximation provided by parameters::p_value_kenward().
This computation can take a while. The confidence-interval calculation uses
default confidence interval calculation method used by
broom.mixed::tidy.merMod().
model <- lm(breaks ~ wool * tension, warpbreaks) # default to: b (beta), e (error), s (statistic), p (p value) fmt_effect_md(model, "woolB", "besp") fmt_effect_md(model, "woolB", "Besp", b_lab = "WoolB") fmt_effect_md(model, "woolB", "i")model <- lm(breaks ~ wool * tension, warpbreaks) # default to: b (beta), e (error), s (statistic), p (p value) fmt_effect_md(model, "woolB", "besp") fmt_effect_md(model, "woolB", "Besp", b_lab = "WoolB") fmt_effect_md(model, "woolB", "i")
Format a number with a fixed number of digits
fmt_fix_digits(xs, digits = 2)fmt_fix_digits(xs, digits = 2)
xs |
a vector of numbers or a character vector representing numbers |
digits |
number of digits of precision |
# what we want to avoid as.character(round(c(.4001, .1000, .5500), 2)) fmt_fix_digits(c(.4001, .1000, .5500), 1) fmt_fix_digits(c(.4001, .1000, .5500), 2) fmt_fix_digits(c(.4001, .1000, .5500), 3)# what we want to avoid as.character(round(c(.4001, .1000, .5500), 2)) fmt_fix_digits(c(.4001, .1000, .5500), 1) fmt_fix_digits(c(.4001, .1000, .5500), 2) fmt_fix_digits(c(.4001, .1000, .5500), 3)
Format numbers to remove leading zeros
fmt_leading_zero(xs)fmt_leading_zero(xs)
xs |
a vector of numbers or a character vector representing numbers |
APA format says that values that are bounded between [-1, 1] should not be formatted with a leading zero. Common examples would be correlations, proportions, probabilities and p-values. Why print the digit if it's almost never used?
Zeros are printed to match the precision of the most precise number. For
example, c(0, 0.111) becomes c(.000, .111)
the vector with leading zeros removed. This function returns a warning if any of the values have an absolute value greater than 1.
fmt_leading_zero(c(0, 0.111)) fmt_leading_zero(c(0.99, -0.9, -0.0))fmt_leading_zero(c(0, 0.111)) fmt_leading_zero(c(0.99, -0.9, -0.0))
Format negative numbers with a minus sign
fmt_minus_sign(xs)fmt_minus_sign(xs)
xs |
a vector of numbers or a character vector representing numbers |
Negative zero -0, which might happen from aggressive rounding,
does not get a minus sign.
the vector with leading hyphens replaced with HTML minus signs
(−).
fmt_minus_sign(c(1, .2, -1, -.2)) # Don't allow zero to be signed fmt_minus_sign(c(-0, round(-0.001)))fmt_minus_sign(c(1, .2, -1, -.2)) # Don't allow zero to be signed fmt_minus_sign(c(-0, round(-0.001)))
Format a p-value
fmt_p_value(xs, digits = 3)fmt_p_value(xs, digits = 3)
xs |
a vector of numbers or a character vector representing numbers |
digits |
number of digits of precision |
formatted *-values. Values smaller than the precision 1 / (10 ^ digits) are replaced with a less than statement < [precision].
p <- c(1, 0.1, 0.01, 0.001, 0.0001) fmt_p_value(p, digits = 2) fmt_p_value(p, digits = 3)p <- c(1, 0.1, 0.01, 0.001, 0.0001) fmt_p_value(p, digits = 2) fmt_p_value(p, digits = 3)
Format a p-value in markdown
fmt_p_value_md(ps)fmt_p_value_md(ps)
ps |
p-values to format |
fmt_p_value() is for formatting p-values with manual precision, but this
functions follows some reasonable defaults and returns a markdown formatted
string.
Values less than .06 are formatted with 3 digits. Values equal to .06 or greater are formatted with 2 digits.
scales::label_pvalue() does the initial rounding and formatting. Then this
function strips off the leading 0 of the p value.
a character vector of markdown formatted p-values
fmt_p_value_md(0.0912) fmt_p_value_md(0.0512) fmt_p_value_md(0.005) # "p less than" notation kicks in below .001. fmt_p_value_md(0.0005)fmt_p_value_md(0.0912) fmt_p_value_md(0.0512) fmt_p_value_md(0.005) # "p less than" notation kicks in below .001. fmt_p_value_md(0.0005)
Replace HTML entities used by this package with UTF-8 codes
fmt_remove_html_entities(xs)fmt_remove_html_entities(xs)
xs |
a character vector |
the updated character vector
x <- "a < −12" |> fmt_remove_html_entities() x charToRaw(x) charToRaw("a < -12") fmt_remove_html_entities("1–2")x <- "a < −12" |> fmt_remove_html_entities() x charToRaw(x) charToRaw("a < -12") fmt_remove_html_entities("1–2")
Replace NAs with another value
fmt_replace_na(xs, replacement = "")fmt_replace_na(xs, replacement = "")
x |
a character vector |
the updated vector
skel_conf_interval() is a vectorized function. Use it to make multiple
intervals from, say, data-frame columns. skel_conf_interval_pair() is the
unvectorized function. Use it to make a single interval from a vector (pair) of two
numbers.
skel_conf_interval(xs, ys, skeleton = "[{xs}, {ys}]") skel_conf_interval_pair(x, skeleton = "[{x[1]}, {x[2]}]")skel_conf_interval(xs, ys, skeleton = "[{xs}, {ys}]") skel_conf_interval_pair(x, skeleton = "[{x[1]}, {x[2]}]")
xs |
a vector of the first elements in the intervals |
ys |
a vector of the second elements in the intervals |
skeleton |
glue-style format to fill. defaults to |
x |
a vector of two elements to plug into the confidence interval |
These functions are wrappers around calls to glue::glue().
Originally, skel_conf_interval() was named skel_conf_interval_v().
strings representing confidence intervals
skel_conf_interval(c(.1, .2), c(.3, .4)) skel_conf_interval_pair(c(.1, .3))skel_conf_interval(c(.1, .2), c(.3, .4)) skel_conf_interval_pair(c(.1, .3))
skel_range() is a vectorized function. Use it to make multiple range from,
say, data-frame columns. skel_range_pair() is the unvectorized function.
Use it to make a single range from a vector (pair) of two numbers.
skel_range() is a vectorized function. Use it to make multiple range from,
say, data-frame columns. skel_range_pair() is the unvectorized function.
Use it to make a single range from a vector (pair) of two numbers.
skel_range_pair(x, skeleton = "{x[1]}–{x[2]}") skel_range(xs, ys, skeleton = "{xs}–{ys}")skel_range_pair(x, skeleton = "{x[1]}–{x[2]}") skel_range(xs, ys, skeleton = "{xs}–{ys}")
x |
a vector of two elements to plug into the range |
skeleton |
glue-style format to fill. defaults to |
xs |
a vector of the first elements in the range |
ys |
a vector of the second elements in the range |
These functions are wrappers around calls to glue::glue().
These functions are wrappers around calls to glue::glue().
strings representing ranges
strings representing ranges
skel_range(c(.1, .2), c(.3, .4)) skel_range_pair(c(.1, .3)) skel_range(c(.1, .2), c(.3, .4)) skel_range_pair(c(.1, .3))skel_range(c(.1, .2), c(.3, .4)) skel_range_pair(c(.1, .3)) skel_range(c(.1, .2), c(.3, .4)) skel_range_pair(c(.1, .3))
Skeletons for inline stats
skel_se(x, skeleton = "SE = {x}") skel_ci(x, ci_width = "95", skeleton = "{ci_width}% CI = {x}")skel_se(x, skeleton = "SE = {x}") skel_ci(x, ci_width = "95", skeleton = "{ci_width}% CI = {x}")
skeleton |
glue-style format to fill. defaults to |
ci_width |
width of the confidence interval to report. Defaults to
|
xs |
a vector of the values to plug into the skeleton |
strings with stats plugged in.
This skeleton handles formats like t-statistics (t(df) = value) or
correlations (r(df) = value).
skel_stat_n_value_pair( x, stat = "t", skeleton = "{stat}({x[1]}) = {x[2]}" )skel_stat_n_value_pair( x, stat = "t", skeleton = "{stat}({x[1]}) = {x[2]}" )
x |
a two-element vector where the first number is the argument to the statistical function and the second is its value. |
stat |
symbol for the statistic. defaults to |
skeleton |
glue-style format to fill. defaults to
|
the formatted string
The common use of this function to clean up columns in a presentation-quality table.
str_replace_same_as_previous(string, replacement)str_replace_same_as_previous(string, replacement)
string |
a character vector |
replacement |
text to use as a replacement for duplicated values |
a single character vector with immediately repeating items replaced
str_replace_same_as_previous( c("a", "a", "a", "b", "b", "c", "d", "d"), "" )str_replace_same_as_previous( c("a", "a", "a", "b", "b", "c", "d", "d"), "" )
The usual job of this function is to break a string into a vector of individual characters, but it can break strings using other separators.
str_tokenize(string, pattern = NULL)str_tokenize(string, pattern = NULL)
string |
a character vector of strings to break |
pattern |
pattern to use for splitting. Defaults to |
a single character vector of the tokens
str_tokenize(c("abc", "de")) str_tokenize(c("abc de fg"), " ")str_tokenize(c("abc", "de")) str_tokenize(c("abc de fg"), " ")
This function is a streamlined, recursive version of
split().
super_split(.data, ...)super_split(.data, ...)
.data |
a dataframe |
... |
(unquoted) names of columns to split by |
a list of dataframes when splitting by a single variable, a list of lists of dataframes when splitting by 2 variables, and so on.
# some kind of 2 by 2 design df <- data.frame( x = c(1, 2, 3, 4, 5, 6, 7, 8), time = c(1, 1, 2, 2, 1, 1, 2, 2), group = c("a", "a", "a", "a", "b", "b", "b", "b") ) super_split(df, group) super_split(df, time) # split by group and then split each of those by time super_split(df, group, time)# some kind of 2 by 2 design df <- data.frame( x = c(1, 2, 3, 4, 5, 6, 7, 8), time = c(1, 1, 2, 2, 1, 1, 2, 2), group = c("a", "a", "a", "a", "b", "b", "b", "b") ) super_split(df, group) super_split(df, time) # split by group and then split each of those by time super_split(df, group, time)