In XikunHan/quickReg: Build Regression Models and Display the Results Using 'ggplot2'

A manual to show the R package quickReg.

Introduction

The quickReg package provide a set of functions to display and pry a dataset. More precisely, the package can display statistical descriptions for a dataset, build regression models for lm, glm and cox regressions based on specified independent variables. The package also provides several seamless functions to visualize the regression results. Some examples are shown below.

Getting started

The example data is a toy dataset extracting from package PredictABEL only to be used to demonstrate the main idea of the package.

# download it from cran or github

# install.packages("quickReg")

# devtools::install_github("XikunHan/quickReg", build_vignettes = TRUE)

library(quickReg)
library(ggplot2)
library(rlang)
library(dplyr)

# Load the dataset

data(diabetes)

# Show the first 6 rows of the data

head(diabetes)

Functions display_table or display_table_group can be used to show statistical descriptions of the dataset.

display_1<-display_table(data=diabetes,variables=c("age","smoking","education"),group="CFHrs2230199")
display_1

# sub-group analysis by sex
display_2<-display_table_group(data=diabetes,variables=c("age","smoking"),group="CFHrs2230199",super_group = "sex")
display_2

# sub-group analysis by two variables
display_3<-display_table_group(data=diabetes,variables=c("age","smoking"),group="CFHrs2230199",super_group = c("sex","education"))
display_3


# Sub-group analysis can be a combination
display_4<-display_table_group(data=diabetes,variables=c("age","smoking"),group="CFHrs2230199",super_group = c("sex","education"),group_combine = TRUE)
display_4

Build regression models

# Apply univariate regression models

reg_1<-reg_x(data = diabetes, y = 5, factors = c(1, 3, 4), model = 'glm')
reg_1

# Or survival analysis

reg_2<-reg_x(data = diabetes, x = c(3:4, 6), y ="diabetes",time=2,factors = c(1, 3, 4), model = 'coxph')
reg_2

# adjusting for covariates

reg_3<-reg_x(data = diabetes, x = c("sex","age"), y ="diabetes" ,cov=c("CFBrs641153","CFHrs2230199"), factors ="sex", model = 'glm',cov_show = TRUE)
reg_3


# regression on several dependent variables
reg_4<-reg_y(data = diabetes, x = c("sex","age","CFHrs1061170"), y =c("systolic","diastolic","BMI") ,cov=c("CFBrs641153","CFHrs2230199"), factors ="sex", model = 'lm')
reg_4

# do a subgroup analysis

reg_5<-reg(data = diabetes, x = c("age","CFHrs1061170"), y =c("systolic","diastolic") ,cov=c("CFBrs641153","CFHrs2230199"), model = 'lm',group="sex")
reg_5


# or two subgroup analysis
reg_6<-reg(data = diabetes, x = c("age","CFHrs1061170"), y =c("systolic","diastolic") ,cov=c("CFBrs641153","CFHrs2230199"), model = 'lm',group=c("sex","smoking"))
reg_6


# or subgroup combination analysis
reg_7<-reg(data = diabetes, x = c("age","CFHrs1061170"), y =c("systolic","diastolic") ,cov=c("CFBrs641153","CFHrs2230199"), model = 'lm',group=c("sex","smoking"),group_combine = TRUE)
reg_7

Plot these regression models

plot(reg_1)

# if the OR value is very large, we can set a limit
plot(reg_1,limits=c(NA,3))


# Sort the variables alphabetically

plot(reg_1,limits=c(NA,3), sort ="alphabetical")

# Similarly, we can plot several dependent variables result

plot(reg_4)


# Subgroup and several dependent variables
plot(reg_5)+facet_grid(sex~y)



# modify the plot 
library(ggplot2);library(ggthemes)

plot(reg_1,limits=c(0.5,2))+
  labs(list(title = "Regression Model", x = "variables"))+
  theme_classic() %+replace% 
  theme(legend.position ="none",axis.text.x=element_text(angle=45,size=rel(1.5)))

XikunHan/quickReg documentation built on March 21, 2022, 3:29 p.m.