gge_coef: Dotwhisker plot

In DiogoFerrari/edar: Exploratory Data Analysis in R

Dotwhisker plot

Description

This is a wrap function for the function dwplot from the package dotwhisker

Usage

gge_coef(
  model,
  xlab = "Coefficient Estimate",
  ylab = "",
  var.order = NULL,
  model.id = NULL,
  title = NULL,
  subtitle = NULL,
  footnote = NULL,
  legend.position = "top",
  legend.title = "",
  title.position = "left",
  hc = FALSE,
  hc.type = c("hc3", "hc0", "hc1", "hc2", "hc4")
)

Arguments

model	either an object from lm, glm, or a nnet::multinom or a tidy data.frame with the model summary. If the data frame is provided, it must contain a column named term (the name of the covariates), estimate (the point estimate), and either std.error (standard errors of the estimates) or conf.low or conf.high (low and high intervals of the point estimate). If there are more than one model in the data.frame, it must contain a column with labels identifying each model.
xlab	string with text to display in the x-axis
ylab	string with text to display in the y-axis
var.order	A vector of variable names that specifies the order in which the variables are to appear along the y-axis of the plot. If a data.frame is provided, the order of the y-axis will follow the order of the rows in the data.frame
model.id	a string with the name of the column that contain the id of each model
title	a string, the title of the plot
subtitle	a string, the subtitle of the plot
footnote	a string, the footnote of the plot
legend.position	a string (top, bottom, left (Default)), right
legend.title	a string with the title of the legend
title.position	a string (or an number) with center (or .5), left (or 0), or right (or 1)
hc	boolean, if TRUE robust standard errors (heteroskedasticity corrected) are returned
hc.type	a string with the method to compute the standard errors (see hccm)

Details

Robust standard errors are computed when hc=T only when the parameter model is an object from lm, glm.

Examples

library(magrittr)
set.seed(77)

data = tibble::data_frame(n = 300,
                          x1   = rnorm(n,3,1),
                          x2   = rexp(n),
                          cat1 = sample(c(0,1), n, replace=TRUE),
                          cat2 = sample(letters[1:4], n, replace=TRUE),
                          y    = -10*x1*cat1 + 10*x2*(3*(cat2=='a')
                                 -3*(cat2=='b') +1*(cat2=='c') -1*(cat2=='d')) + 
                              rnorm(n,0,10), 
                          y.bin = ifelse(y < mean(y), 0, 1),
                          y.mul = 1+ifelse( - x1 - x2 + rnorm(n,sd=10) < 0, 0,
                                    ifelse( - 2*x2 + rnorm(n,sd=10) < 0, 1, 2)),
                          )

model.g1 = lm(y ~ x1, data)
model.g2 = lm(y ~ x1 + x2, data)
model.g  = lm(y ~ x1*cat1 + x2*cat2, data)
model.bin = glm(y.bin ~ x1+x2*cat2, data=data, family='binomial')
model.mul <- nnet::multinom(y.mul ~ x1 + x2, data)


gge_coef(model.g)
gge_coef(model.g, hc=TRUE)
gge_coef(tidye(model.g) %>% dplyr::arrange(estimate), hc=TRUE)

gge_coef(model.bin)
gge_coef(model.bin, hc=TRUE)
gge_coef(tidye(model.g))

## many models at once
models=tidye(list('Standard Model'=model.bin)) %>%
    dplyr::bind_rows(tidye(list('Robust std. error'=model.bin), hc=TRUE) )
gge_coef(models)
gge_coef(models, model.id='model')

## list of models
gge_coef(list(model.g, model.g1, model.g2), model.id='model')
gge_coef(list('Complete model'=model.g, "One Covar"=model.g1,
              "Two covars"=model.g2), model.id='model' )

## pipe can be used
## modelg %>% gge_coef

DiogoFerrari/edar documentation built on May 8, 2022, 8:26 a.m.