5.2 Formatting

5.2.1 Post Processing

By default, Tplyr will build your table and store it as a tibble. It can be accessed directly in your environment by simply calling it.

head(my_table, n = 9)
## # A tibble: 9 × 9
##   row_label1          row_label2 var1_Placebo     var1_Total    `var1_Xanomeli…`
##   <chr>               <chr>      <chr>            <chr>         <chr>           
## 1 Age (years)         N          "86"             "254"         "84"            
## 2 Age (years)         Mean [SD]  "75.21 [ 8.590]" "75.09 [ 8.2… "74.38 [ 7.886]"
## 3 Age (years)         Median     "76.0"           "77.0"        "76.0"          
## 4 Age (years)         Min, Max   "52, 89"         "51, 89"      "56, 88"        
## 5 Age Group 1 (years) <65        " 14 [16.28%]"   " 33 [12.99%… " 11 [13.10%]"  
## 6 Age Group 1 (years) >80        " 30 [34.88%]"   " 77 [30.31%… " 18 [21.43%]"  
## 7 Age Group 1 (years) 65-80      " 42 [48.84%]"   "144 [56.69%… " 55 [65.48%]"  
## 8 Gender              F          " 53 [61.63%]"   "143 [56.30%… " 40 [47.62%]"  
## 9 Gender              M          " 33 [38.37%]"   "111 [43.70%… " 44 [52.38%]"  
## # … with 4 more variables: `var1_Xanomeline Low Dose` <chr>,
## #   ord_layer_index <int>, ord_layer_1 <int>, ord_layer_2 <dbl>

Since it is tibble, you are free to further process or wrangle it to your liking. Below we’ll apply some light formatting using a combination of Tplyr and dplyr:

  1. Remove repeating variable labels and insert breaks between variable rows
  2. Reorder the columns
  3. Remove extraneous columns not of interest for display (i.e. meta data created by Tplyr)
library(dplyr)

my_table <- my_table %>%
  
  # remove repeating labels
  apply_row_masks(., row_breaks = TRUE) %>%
  
  # specify order of relevant variables
  select(row_label1, 
         row_label2,
         `var1_Xanomeline High Dose`,
         `var1_Xanomeline Low Dose`,
         var1_Placebo,
         var1_Total)

We can see the formatting has taken effect by recalling it in our environment. Specifically, compare the row_label1 variable below to the previous section while also noting the order of columns has changed to our specification.

head(my_table, n = 9)
## # A tibble: 9 × 6
##   row_label1           row_label2 `var1_Xanomeli…` `var1_Xanomeli…` var1_Placebo
##   <chr>                <chr>      <chr>            <chr>            <chr>       
## 1 "Age (years)"        "N"        "84"             "84"             "86"        
## 2 ""                   "Mean [SD… "74.38 [ 7.886]" "75.67 [ 8.286]" "75.21 [ 8.…
## 3 ""                   "Median"   "76.0"           "77.5"           "76.0"      
## 4 ""                   "Min, Max" "56, 88"         "51, 88"         "52, 89"    
## 5 ""                   ""         ""               ""               ""          
## 6 "Age Group 1 (years… "<65"      " 11 [13.10%]"   "  8 [ 9.52%]"   " 14 [16.28…
## 7 ""                   ">80"      " 18 [21.43%]"   " 29 [34.52%]"   " 30 [34.88…
## 8 ""                   "65-80"    " 55 [65.48%]"   " 47 [55.95%]"   " 42 [48.84…
## 9 ""                   ""         ""               ""               ""          
## # … with 1 more variable: var1_Total <chr>

5.2.2 Displaying

While Tplyr aided us in building a summary tables, other packages exist to help us display them. One package we will focus on is the flextable package. This table display package is quite powerful and offers a high degree of customization.

We turn the interested reader to the flextable book to view additional examples.

library(flextable)

# a basic flextable
my_flextable <- my_table %>%
  
  # start flextable 
  flextable() %>%
  autofit()
my_flextable

We can apply some flextable formatting to alter the display of the table.

# a nicer flextable
my_flextable <- my_table %>%
  
  # start flextable 
  flextable() %>% 
  autofit() %>%
  
  # add some padding between rows
  padding(padding = 0.5) %>%
  
  # adjust width of first two columns
  width(j = 1:2, width = 4) %>%
  
  # align treatment columns to center
  align(part = "all", align = "center", j = 3:6) %>%
  
  # column header labels
  set_header_labels(., values = list(
    row_label1 = 'Variable',
    row_label2 = ' ',
    `var1_Xanomeline High Dose` = 'Xanomeline \nHigh Dose',
    `var1_Xanomeline Low Dose` = 'Xanomeline \nLow Dose',
    var1_Placebo = 'Placebo',
    var1_Total = 'Total')) %>%
  
  # header + footers
  add_header_lines(values = "Table: Demographics (Safety Analysis Set)") %>%
  add_footer_lines(values = "This was produced in R!") %>%
  
  # font size, font name
  fontsize(part = "all", size = 8)
  # font()
  # font(part = "all", fontname = "Times")  
  # This errors, perhaps version issue.  -- SZ