GimmeMyStats: Statistics Utilities

Description

Facilitate reporting for regression and correlation modeling, hypothesis testing, variance analysis, outlier detection, and detailed descriptive statistics.

Author(s)

Maintainer: Etienne Camenen etienne.camenen@gmail.com

See Also

Useful links:

• https://github.com/ecamenen/GimmeMyStats

• https://ecamenen.github.io/GimmeMyStats/

• Report bugs at https://github.com/ecamenen/GimmeMyStats/issues

Add P-value Significance Symbols

Description

Redefine the default parameters of rstatix::add_significance() by adding p-value significance symbols to a data frame.

Usage

add_significance0(data, p.col = NULL, output.col = NULL)

Arguments

Value

a data frame

Examples

library(magrittr)
library(rstatix, warn.conflicts = FALSE)
data("ToothGrowth")
ToothGrowth %>%
    t_test(len ~ dose) %>%
    adjust_pvalue() %>%
    add_significance0("p.adj")

Frequency of categorical variables

Description

Formats a data frame or vector containing categorical variables and calculates the frequency of each category.

Usage

count_category(x, width = 15, collapse = FALSE, sort = TRUE, format = TRUE)

Arguments

Value

A tibble with one row per category and the following columns:

f

Factor specifying the category labels, possibly wrapped to the specified width. When collapse = TRUE, multiple categories with identical frequencies are merged into a single label separated by commas.

n

Integer specifying the frequency count for each category.

Examples

# Vector of categorical variable
k <- 5
n <- runif(k, 1, 10) %>% round()
x <- paste("Level", seq(k)) %>%
    mapply(function(x, y) rep(x, y), ., n) %>%
    unlist()
count_category(x)

# Data frame of categorical variable
df <- sapply(seq(k), function(x) runif(10) %>% round()) %>% as.data.frame()
colnames(df) <- paste("Level", seq(k))
count_category(df)
count_category(x, sort = FALSE, width = 5)
count_category(x, sort = seq(k), format = FALSE)
x2 <- c(x, rep("Level 6", n[1]))
count_category(x2, collapse = TRUE)

Household tasks distribution by gender and arrangement

Description

A dataset containing the distribution of household tasks among different arrangements: Wife, Alternating, Husband, and Jointly. The data represents the frequency of each task performed by each arrangement.

Usage

data(housetasks)

Format

A data.frame with 13 rows (tasks) and 4 columns (arrangements):

Wife

Numeric, the frequency of the task performed primarily by the wife.

Alternating

Numeric, the frequency of the task performed in an alternating manner.

Husband

Numeric, the frequency of the task performed primarily by the husband.

Jointly

Numeric, the frequency of the task performed jointly by both partners.

Source

The dataset was downloaded from the ggpubr GitHub repository: https://raw.githubusercontent.com/kassambara/ggpubr/refs/heads/master/inst/demo-data/housetasks.txt

Examples

data(housetasks)
head(housetasks)

Identifies outliers in a numeric vector

Description

Detects outliers using methods like IQR, percentiles, Hampel, MAD, or SD.

Usage

identify_outliers(
  x,
  probabilities = c(0.25, 0.75),
  method = "iqr",
  weight = 1.5,
  replace = FALSE
)

Arguments

Value

A numeric vector whose content depends on the value of replace:

replace = FALSE

A numeric vector containing only the detected outlier values. The vector is named with the original indices or names of x.

replace = TRUE

A numeric vector of the same length as x, where detected outliers are replaced by NA.

Examples

x <- rnorm(100)
identify_outliers(x, method = "iqr")
identify_outliers(x, method = "percentiles", probabilities = c(0.1, 0.9))
identify_outliers(x, method = "sd", weight = 3)
identify_outliers(x, method = "mad", replace = TRUE)

Multiple correlation test

Description

Calculates correlations between multiple variables.

Usage

mcor_test(
  x,
  y = NULL,
  estimate = TRUE,
  p.value = FALSE,
  method = "spearman",
  method_adjust = "BH"
)

Arguments

Value

Depending on the values of estimate and p.value, one of the following:

estimate = TRUE, p.value = FALSE

A numeric matrix of correlation coefficients, with columns corresponding to variables in x and rows to variables in y.

estimate = FALSE, p.value = TRUE

A numeric matrix of adjusted p-values, with columns corresponding to variables in x and rows to variables in y.

estimate = TRUE, p.value = TRUE

A named list with two elements:

estimate

Numeric matrix of correlation coefficients.

p.value

Numeric matrix of adjusted p-values.

Examples

library(magrittr)
x0 <- runif(20)
x <- lapply(
    c(1, -1),
    function(i) sapply(seq(10), function(j) x0 * i + runif(10, max = 1))
) %>%
    Reduce(cbind, .) %>%
    set_colnames(paste("Variable", seq(20)))
y <- lapply(
    c(1, -1),
    function(i) sapply(seq(10), function(j) x0 * i + runif(10, max = 1))
) %>%
    Reduce(cbind, .) %>%
    set_colnames(paste("Variable", seq(20))) %>%
    .[, seq(5)]
mcor_test(x)
mcor_test(
    x,
    y,
    p.value = TRUE,
    method = "pearson",
    method_adjust = "bonferroni"
)

Performs post hoc analysis for chi-squared or Fisher's exact test

Description

Identifies pairwise differences between categories following a chi-squared or Fisher's exact test.

Usage

post_hoc_chi2(
  x,
  method = "fisher",
  method_adjust = "BH",
  digits = 3,
  count = FALSE,
  ...
)

Arguments

Details

If x is numeric, it is treated as a contingency table and the names are considered as categories; otherwise, the levels of the factor or the characters are used.

Value

A tibble with pairwise test results containing the following columns:

group1, group2

Character vectors specifying the pair of groups being compared.

n

Numeric vector specifying the total count or sample size for the comparison.

statistic

Numeric vector specifying the test statistic (for chi-squared tests only).

df

Numeric vector specifying the degrees of freedom (for chi-squared tests only).

p

Raw p-value for the pairwise comparison, formatted as numeric or character ("< 0.001" for very small p-values).

p.signif

Character vectors specifying the significance codes for raw p-values: 'ns' (not significant).

FDR

False Discovery Rate adjusted p-value using the specified method, formatted as numeric or character ("< 0.001" for very small values).

fdr.signif

Character vectors specifying the significance codes for FDR-adjusted p-values: 'ns' (not significant), '' (p < 0.05), '' (p < 0.01), '' (p < 0.001).

For Fisher's exact tests, the statistic and df columns are not included..

Examples

x <- c(rep("A", 100), rep("B", 78), rep("C", 25))
post_hoc_chi2(x)

x <- data.frame(G1 = c(Yes = 100, No = 78), G2 = c(Yes = 75, No = 23))
post_hoc_chi2(x, count = TRUE, method = "chisq")

data("housetasks")
housetasks[, c("Wife", "Husband")] %>%
    t() %>%
    post_hoc_chi2(count = TRUE, workspace = 1e6)

x <- cbind(
    mapply(function(x, y) rep(x, y), letters[seq(3)], c(7, 5, 8)) %>% unlist(),
    mapply(function(x, y) rep(x, y), LETTERS[seq(3)], c(6, 6, 8)) %>% unlist()
)
post_hoc_chi2(x)

Prints descriptive statistics for binomial variables

Description

Calculates and prints frequency counts and percentages for binomial (two-level) categorical variables.

Usage

print_binomial(x, digits = 1, width = 15)

Arguments

Value

A tibble with one row per level for each categorical level containing the following columns:

Variables

Character vector specifying the name of each variable.

Levels

Character vector specifying the category level for each variable.

Statistics

Character vector combining the frequency count and the percentage for each level.

Examples

x <- data.frame(A = sample(c("X", "Y"), 100, replace = TRUE))
print_binomial(x)
print_binomial(x, digits = 2, width = 5)

Prints the results of a Chi2

Description

Formats the results of a Chi-squared or Fisher's exact test.

Usage

print_chi2_test(x, digits = 3)

Arguments

Value

A character string containing the formatted test results with:

Test statistic

For Chi-squared test.

P-value

Formatted p-value with significance stars.

Sample size

Total count for sample size.

For Fisher's exact test, only the P-value and sample size are included.

Examples

x <- c(A = 100, B = 78, C = 25)
library(rstatix)
print_chi2_test(chisq_test(x))

xtab <- as.table(rbind(c(490, 10), c(400, 100)))
dimnames(xtab) <- list(
    group = c("grp1", "grp2"),
    smoker = c("yes", "no")
)
print_chi2_test(fisher_test(xtab))

Prints the dispersion of a numeric vector

Description

Calculates and prints the median and interquartile range (IQR) or the mean and standard deviation (SD).

Usage

print_dispersion(x, digits = 1, width = 15, method = "median")

Arguments

Value

A character string containing a measure of central tendency and dispersion. Depending on method, this is either the median and interquartile range or the mean and standard deviation.

Examples

print_dispersion(runif(10))
print_dispersion(runif(10), method = "mean", digits = 2, width = 5)

Prints descriptive statistics for multinomial variables

Description

Calculates and prints frequency counts and percentages for multinomial (multi-level) categorical variables.

Usage

print_multinomial(x, label = NULL, digits = 1, width = 15, n = nrow(x), ...)

Arguments

Value

A tibble with one row per level for each categorical level containing the following columns:

Variables

Character vector specifying the name of each variable.

Levels

Character vector specifying the category level for each variable.

Statistics

Character vector combining the frequency count and the percentage for each level.

Examples

x <- data.frame(A = sample(c("X", "Y", "Z"), 100, replace = TRUE))
print_multinomial(x, label = "A")
x2 <- rbind(x, data.frame(A = rep("Level A", length(x[x == "Level X", ]))))
print_multinomial(
    x,
    label = "Variable A",
    sort = FALSE,
    n = 90,
    digits = 2,
    width = 5
)

Prints descriptive statistics for numeric variables

Description

Prints summary statistics (mean, median, quartiles, range, etc.) for numeric variables.

Usage

print_numeric(x, digits = 1, width = 15)

Arguments

Value

A tibble with one row per numeric variable and the following columns:

Variables

Character specifying the variable name.

Mean+/-SD

Character specifying the mean and standard deviation.

Median+/-IQR

Character specifying the median and interquartile range.

Q1-Q3

Character specifying the first and third quartiles.

Range

Character specifying the minimum and maximum values.

Kurtosis

Numeric specifying the kurtosis coefficient.

Skewness

Numeric specifying the skewness coefficient.

Normality

Character specifying the Shapiro-Wilk normality test significance code.

Zeros

Integer specifying the number of zero values.

NAs

Integer specifying the number of missing values.

Examples

x <- data.frame(A = rnorm(100), B = rnorm(100))
print_numeric(x)
print_numeric(x, digits = 2, width = 5)

Prints a hypothesis test

Description

Formats the results of a hypothesis test (ANOVA, Kruskal-Wallis, or Wilcoxon).

Usage

print_test(x, digits = 0, digits_p = 2)

Arguments

Value

A character string containing the formatted test results with:

Test name

Name of the statistical test (ANOVA, Kruskal-Wallis, Wilcoxon, t-test, Friedman, or mixed-effects model).

Test statistic

Test statistic (F, K, W, T, or \chi^2) with degrees of freedom when applicable.

P-value

P-value with significance stars.

Examples

library(rstatix)
data("ToothGrowth")
res <- anova_test(ToothGrowth, len ~ dose)
print_test(res)

res <- kruskal_test(ToothGrowth, len ~ dose)
print_test(res)

res <- wilcox_test(ToothGrowth, len ~ supp)
print_test(res)

library(lmerTest)
data("sleepstudy", package = "lme4")
res <- lmer(Reaction ~ Days + (Days | Subject), sleepstudy)
print_test(res)

Summarizes descriptive statistics for binomial variables

Description

Summarizes descriptive statistics for binomial variables

Usage

summary_binomial(x, ...)

Arguments

Value

A tibble with descriptive statistics containing the following columns:

Variables

Character vector specifying the name of each variable.

Statistics

Character vector combining the reference level of a variable with its frequency count and its percentage.

Examples

x <- data.frame(A = sample(c("X", "Y"), 100, replace = TRUE))
summary_binomial(x)
summary_binomial(x, digits = 2, width = 5)

Summarizes descriptive statistics for numeric variables

Description

Formats the output of print_numeric into a concise summary.

Usage

summary_numeric(x, ...)

Arguments

Value

A tibble with one row per numeric variable and the following columns:

Variables

Character specifying the variable name.

Median+/-IQR

Character specifying the median and interquartile range.

Examples

x <- data.frame(A = rnorm(100), B = rnorm(100))
summary_numeric(x)
summary_numeric(x, digits = 2, width = 5)

Convert Strings to Title Case

Description

Converts the first character of each string to uppercase and the rest to lowercase.

Usage

to_title(x)

Arguments

Value

A character vector with the same length as x, where each element has its first character converted to uppercase and remaining characters are preserved as-is.

Examples

to_title(c("hELLO", "WoRLD", "R"))
# Returns: "Hello" "World" "R"