01_table1_demographic.Rmd

---
title: "01_table1_demographic"
author: "randy"
date: "`r Sys.Date()`"
output: pdf_document
---

::: callout-important
-   Table1 has three different versions :
    -   the basic training & testing

    -   the basic training & testing + overall

    -   the basic training & testing + pvalues
-   Do we need to include too many variables not used in the paper?
    -   ethnicity

    -   genotype
:::

## Setup

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE,
                  warnings = FALSE,
                  message = FALSE,
                  comment = "#>",
                  #results = "hide",
                  error = FALSE)

## clean the R environment
graphics.off()
rm(list = ls())
freshr::freshr()

## load packages
library(here, quietly = TRUE)
library(tidyverse, quietly = TRUE)
library(gtsummary, quietly = TRUE)
library(flextable, quietly = TRUE)

## check the directory for the file
# here::dr_here()
here::set_here()

## the figure or results should be saved 
# paste0("foldername/Sfilename_workingresult_", 
#.       Sys.Date(), ".filetype")
```



```{r "dataset", eval = FALSE}
## set seed
set.seed(555)
# load("data/sysdata.rda")
load("final/epic_clean_full_data.Rdata")

# the code to prepare for
# epic, demog, test and train
# data0, data1, and data2
# they are all saved in sysdata.rda files
data <- left_join(epic, demog, by = "id") %>%
  mutate(sex = as.factor(sex))
test_id <- unique(test$id) %>% unlist()
train_id <- unique(train$id) %>% unlist()

data0 <- data %>%
  mutate(group = case_when(id %in% test_id ~ "testing",
                           TRUE ~ "training"))
data1 <- data0 %>%
  group_by(id, group) %>%
  summarize(age_mean = mean(age),
            age_min = min(age),
            age_max = max(age),
            age_n = length(age),
            visitn = n(),
            h_mean = mean(ht),
            h_max = max(ht),
            h_min = min(ht),
            w_mean = mean(wt),
            w_max = max(wt),
            w_min = min(wt),
            sex = sex,
            genotype = genotype,
            ethnic = ethnic,
            race = race) %>%
  ungroup() %>%
  unique()

# working dataset
data2 <- full_join(data1, data,
                   by = join_by(id, sex, genotype, ethnic, race)) %>%
  as.data.frame() %>%
  mutate(time = age - age_min,
         age_diff = age_max - age_min)

write.csv(data1, file = paste0("data/S01_table1_dataset_randy_", Sys.Date(), ".csv"))
write.csv(data2, file = paste0("data/S01_epic_clean_randy_", Sys.Date(), ".csv"))
```


## Making table1

```{r "table1", message = FALSE, warning = FALSE}
library(readr)
data1 <- read_csv("data/S01_table1_dataset_randy_2023-08-23.csv")
View(data1)

## table0 contains all the information about demgo for total
table0 <- data1 %>%
  unique() %>%
  dplyr::select(-id) %>%
  mutate(ethnic = case_when(ethnic == 1 ~ "Hispanic",
                            ethnic == 2 ~ "Non-Hispanic"),
         race = case_when(race == 1 ~ "White",
                          race != 1 ~ "Other"),
         sex = case_when(sex == "F" ~ "Female",
                         sex == "M" ~ "Male"),
         age_diff = age_max - age_min) %>%
  dplyr::select(group,
                Genotype = genotype,
                Gender = sex,
                Race = race,
                Ethnicity = ethnic,
                "Visit number" = visitn,
                "Age baseline" = age_min, 
                "Age final" = age_max, 
                "Follow up years" = age_diff,
                # "Height mean" = h_mean, 
                "Height baseline" = h_min) %>% 
  # "Weight mean" = w_mean,
  # "Weight baseline" = w_min
  ## select all the variables for table1
  tbl_summary(by = group) %>%
  ## just display all the variables in one column
  modify_header(label = "**Characteristics**") %>%
  # update the column header
  bold_labels() %>%
  italicize_labels() %>%
  # as.data.frame()
  as_flex_table() %>%
  flextable::bold(part = "header") %>%
  ## auto adjust the column widths
  flextable::autofit()

## table1 contains information of dataset grouped as training and testing
table1 <- data1 %>%
  unique() %>%
  dplyr::select(-id) %>%
  mutate(ethnic = case_when(ethnic == 1 ~ "Hispanic",
                            ethnic == 2 ~ "Non-Hispanic"),
         race = case_when(race == 1 ~ "White",
                          race != 1 ~ "Other"),
         sex = case_when(sex == "F" ~ "Female",
                         sex == "M" ~ "Male"),
         age_diff = age_max - age_min) %>% 
  dplyr::select(group,
                Genotype = genotype,
                Gender = sex,
                Race = race,
                Ethnicity = ethnic,
                "Visit number" = visitn,
                # "Age mean" = age_mean, 
                "Age baseline" = age_min, 
                "Age final" = age_max, 
                "Follow up years" = age_diff,
                # "Height mean" = h_mean, 
                "Height baseline" = h_min) %>% 
  # "Weight mean" = w_mean,
  # "Weight baseline" = w_min)
  ## select all the variables for table1
  tbl_summary(by = group) %>%
  ## just display all the variables in one column
  modify_header(label = "**Characteristics**") %>%
  # update the column header
  bold_labels() %>%
  add_p() %>%
  italicize_labels()


## table1 contains information of dataset grouped as training and testing
table2 <- data1 %>%
  unique() %>%
  dplyr::select(-id) %>%
  mutate(ethnic = case_when(ethnic == 1 ~ "Hispanic",
                            ethnic == 2 ~ "Non-Hispanic"),
         race = case_when(race == 1 ~ "White",
                          race != 1 ~ "Other"),
         sex = case_when(sex == "F" ~ "Female",
                         sex == "M" ~ "Male"),
         age_diff = age_max - age_min) %>% 
  dplyr::select(group,
                Genotype = genotype,
                Gender = sex,
                Race = race,
                Ethnicity = ethnic,
                "Visit number" = visitn,
                # "Age mean" = age_mean, 
                "Age baseline" = age_min, 
                # "Age final" = age_max, 
                "Follow up years" = age_diff,
                # "Height mean" = h_mean, 
                "Height baseline" = h_min) %>% 
  tbl_summary(by = group,
              statistic = list(all_continuous() ~ "{mean} ({sd})") ) %>%
  ## just display all the variables in one column
  modify_header(label = "**Characteristics**") %>%
  # update the column header
  bold_labels() %>%
  add_overall(last = TRUE) %>%
  italicize_labels() 

table0
```

## Saving for the table1

```{r "saving", eval=FALSE, include=FALSE}

## save .pptx file
## flextable can be saved directly to powerpoints
flextable::save_as_pptx(table0, 
                        path = paste0("figure/S01_plmlmm_table1_", Sys.Date(), ".pptx"))


## save the .png file
library(webshot)
# webshot::install_phantomjs()
flextable::save_as_image(table0, 
              path = paste0("figure/S01_plmlmm_table1_", Sys.Date(), ".png"),
              # zoom = 3, 
              # expand = 10, 
              webshot = "webshot")

# flextable::save_as_image(table0, 
#               path = paste0("figure/S01_plmlmm_table0_", Sys.Date(), ".pptx"),
#               # zoom = 3, 
#               # expand = 10, 
#               webshot = "webshot")



## save the latex .text file
library(xtable)
# xtable(table0, type = "latex", 
#        file = paste0("figure/S01_plmlmm_table0_", Sys.Date(), ".tex"))
# print(xtable(table0, type = "latex"),
#       file = paste0("figure/S01_plmlmm_table0_", Sys.Date(), ".tex"))
# xtable(table1, type = "latex", 
#        file = paste0("figure/S01_plmlmm_table1_", Sys.Date(), ".tex"))
print(xtable(table1, type = "latex"),
      file = paste0("figure/S01_plmlmm_table1_pvalue_", Sys.Date(), ".tex"))
# xtable(table2, type = "latex", 
#        file = paste0("figure/S01_plmlmm_table2_", Sys.Date(), ".tex"))
print(xtable(table2, type = "latex"),
      file = paste0("figure/S01_plmlmm_table1_overall_", Sys.Date(), ".tex"))

```

```{r}
sessionInfo()
```