Manipulate and describe a dataset in R

References

https://cran.r-project.org/

https://r4ds.had.co.nz/

https://adv-r.hadley.nz/

https://stats.idre.ucla.edu/r/

Objectives

-Analyze various variables in the birthwt dataset (from the MASS package).

-Evaluate bwt, a continous variable, in relation to other variables.

-Evaluate low, a categorical variable, in relation to other variables.

Libraries

library(tidyverse)

## -- Attaching packages ------------

## v ggplot2 3.2.1     v purrr   0.3.3
## v tibble  2.1.3     v dplyr   0.8.3
## v tidyr   1.0.0     v stringr 1.4.0
## v readr   1.3.1     v forcats 0.4.0

## -- Conflicts ---------------------
## x dplyr::filter() masks stats::filter()
## x dplyr::lag()    masks stats::lag()

library(MASS)

## 
## Attaching package: 'MASS'

## The following object is masked from 'package:dplyr':
## 
##     select

library(psych)

## 
## Attaching package: 'psych'

## The following objects are masked from 'package:ggplot2':
## 
##     %+%, alpha

library(doBy)

library(haven)

Structure of the dataset

data("birthwt")

str(birthwt)

## 'data.frame':    189 obs. of  10 variables:
##  $ low  : int  0 0 0 0 0 0 0 0 0 0 ...
##  $ age  : int  19 33 20 21 18 21 22 17 29 26 ...
##  $ lwt  : int  182 155 105 108 107 124 118 103 123 113 ...
##  $ race : int  2 3 1 1 1 3 1 3 1 1 ...
##  $ smoke: int  0 0 1 1 1 0 0 0 1 1 ...
##  $ ptl  : int  0 0 0 0 0 0 0 0 0 0 ...
##  $ ht   : int  0 0 0 0 0 0 0 0 0 0 ...
##  $ ui   : int  1 0 0 1 1 0 0 0 0 0 ...
##  $ ftv  : int  0 3 1 2 0 0 1 1 1 0 ...
##  $ bwt  : int  2523 2551 2557 2594 2600 2622 2637 2637 2663 2665 ...

Describe the variables

#	variable name	variable label	coded levels
1	low	indicator of birth weight less than 2.5 kg	0, 1
2	age	mother’s age in years	continous variable
3	lwt	mother’s weight in pounds at last menstrual period	continous variable
4	race	mother’s race (1 = white, 2 = black, 3 = other)	1, 2, 3
5	smoke	smoking status during pregnancy	0, 1
6	ptl	number of previous premature labours	0, 1, 2, 3
7	ht	history of hypertension	0, 1
8	ui	presence of uterine irritability	0, 1
9	ftv	number of physician visits during the first trimester	0, 1, 2, 3, 4, 6
10	bwt	birth weight in grams	continous variable

Transform the data.frame

-Factorize -Collapse levels -Create indicator or dummy variables -Labels the levels of a factor

# make a copy of the original data.frame

bwt.df <- birthwt  

names(bwt.df)

##  [1] "low"   "age"   "lwt"   "race"  "smoke" "ptl"   "ht"    "ui"   
##  [9] "ftv"   "bwt"

Make grouping and/or indicator variables for variable race

# Check the levels of the variable

table(bwt.df$race)

## 
##  1  2  3 
## 96 26 67

# No. of levels

table(bwt.df$race) %>% as.data.frame() %>% dplyr::select(Var1) %>% unlist()%>% length() 

## [1] 3

# Levels

table(bwt.df$race) %>% as.data.frame() %>% dplyr::select(Var1) %>% unlist() 

## Var11 Var12 Var13 
##     1     2     3 
## Levels: 1 2 3

# regroup the levels without collapsing

bwt.df$racegr <- 99


bwt.df$racegr[bwt.df$race==1] <- 0
bwt.df$racegr[bwt.df$race==2] <- 1
bwt.df$racegr[bwt.df$race==3] <- 2

table(bwt.df$racegr)

## 
##  0  1  2 
## 96 26 67

table(bwt.df$racegr, bwt.df$race)

##    
##      1  2  3
##   0 96  0  0
##   1  0 26  0
##   2  0  0 67

# create two indicator/dummy variables

bwt.df$black <- 0
bwt.df$other <- 0


bwt.df$black[bwt.df$racegr==1] <- 1
bwt.df$other[bwt.df$racegr==2] <- 1

table(bwt.df$black, bwt.df$other)

##    
##      0  1
##   0 96 67
##   1 26  0

Make grouping and/or indicator variables for variable ptl

# Check levels of the variables

table(bwt.df$ptl)

## 
##   0   1   2   3 
## 159  24   5   1

table(bwt.df$ptl) %>% as.data.frame() %>% dplyr::select(Var1) %>% unlist()%>% length() # No. of levels: 4

## [1] 4

table(bwt.df$ptl) %>% as.data.frame() %>% dplyr::select(Var1) %>% unlist() # Levels: 0 1 2 3

## Var11 Var12 Var13 Var14 
##     0     1     2     3 
## Levels: 0 1 2 3

# collapse into 3 levels


bwt.df$ptlgr <- 99

bwt.df$ptlgr[bwt.df$ptl==0] <- 0
bwt.df$ptlgr[bwt.df$ptl==1] <- 1
bwt.df$ptlgr[bwt.df$ptl==2 | bwt.df$ptl==3] <- 2


table(bwt.df$ptlgr)

## 
##   0   1   2 
## 159  24   6

# create two indicator/dummy variables

bwt.df$one_preterm_labor <- 0
bwt.df$twoOrMore_preterm_labor <- 0


bwt.df$one_preterm_labor[bwt.df$ptlgr==1] <- 1
bwt.df$twoOrMore_preterm_labor[bwt.df$ptlgr==2] <- 1

table(bwt.df$one_preterm_labor)

## 
##   0   1 
## 165  24

table(bwt.df$twoOrMore_preterm_labor)

## 
##   0   1 
## 183   6

Make grouping and/or indicator variables for variable ftv

# check levels

table(bwt.df$ftv)

## 
##   0   1   2   3   4   6 
## 100  47  30   7   4   1

table(bwt.df$ftv) %>% as.data.frame() %>% dplyr::select(Var1) %>% unlist() %>% length() # No. of levels: 6

## [1] 6

table(bwt.df$ftv) %>% as.data.frame() %>% dplyr::select(Var1) %>% unlist() # Levels: 0 1 2 3 4 6

## Var11 Var12 Var13 Var14 Var15 Var16 
##     0     1     2     3     4     6 
## Levels: 0 1 2 3 4 6

# collapse into 3 levels

bwt.df$ftvgr <- 99

bwt.df$ftvgr[bwt.df$ftv==0] <- 0
bwt.df$ftvgr[bwt.df$ftv==1] <- 1
bwt.df$ftvgr[bwt.df$ftv==2 | bwt.df$ftv==3 | bwt.df$ftv==4 | bwt.df$ftv==6] <- 2


table(bwt.df$ftvgr)

## 
##   0   1   2 
## 100  47  42

# create indicator/dummy variables

bwt.df$one_firstTimester_drVisit <- 0
bwt.df$twoOrMore_firstTimester_drVisit <- 0


bwt.df$one_firstTimester_drVisit[bwt.df$ftvgr==1] <- 1
bwt.df$twoOrMore_firstTimester_drVisit[bwt.df$ftvgr==2] <- 1


table(bwt.df$one_firstTimester_drVisit)

## 
##   0   1 
## 142  47

table(bwt.df$twoOrMore_firstTimester_drVisit)

## 
##   0   1 
## 147  42

Factorize variables

bwt.df$lowf <- factor(bwt.df$low) # as.factor would also work

bwt.df$racegf <- factor(bwt.df$racegr)

bwt.df$smokef <- factor(bwt.df$smoke)

bwt.df$ptlgf <- factor(bwt.df$ptlgr)

bwt.df$htf <- factor(bwt.df$ht)

bwt.df$uif <- factor(bwt.df$ui)

bwt.df$ftvgf <- factor(bwt.df$ftvgr)

Create factors with labelled levels

bwt.df$lowfl <- factor(bwt.df$low, levels = c(0, 1), labels = c("High", "Low"))

bwt.df$smokefl <- factor(bwt.df$smoke, levels = c(0, 1), labels = c("Non-smoker", "Smoker"))

bwt.df$htfl <- factor(bwt.df$ht, levels = c(0, 1), labels = c("No_hypertension", "Hypertension"))

bwt.df$uifl <- factor(bwt.df$ui, levels = c(0, 1), labels = c("No_uterine_irritability", "Uterine_irritability"))

bwt.df$racegfl <- factor(bwt.df$racegr, levels = c(0, 1, 2), labels = c("White", "Black", "Other"))

bwt.df$ptlgfl <- factor(bwt.df$ptlgr, levels = c(0, 1, 2), labels = c("No_preterm_labor", "One_preterm_labor", "TwoOrMore_preterm_labor" ))

bwt.df$ftvgfl <- factor(bwt.df$ftvgr, levels = c(0, 1, 2), labels = c("No_physician_visit", "One_firstTimester_drVisit", "twoOrMore_firstTimester_drVisit"))

Rearrange the variables

bwt.df <- 
  
  bwt.df %>% dplyr::select(low, lowf, lowfl, 
                           age, lwt, 
                           race, racegr, racegf, racegfl, black, other, 
                           smoke, smokef, smokefl,
                           ptl, ptlgr, ptlgf, ptlgfl, one_preterm_labor, twoOrMore_preterm_labor,
                           ht, htf, htfl,
                           ui, uif, uifl,
                           ftv, ftvgr, ftvgf, ftvgfl, one_firstTimester_drVisit, twoOrMore_firstTimester_drVisit,
                           bwt
                           )

Structure of the data.frame after manipulations/transformations

str(bwt.df)

## 'data.frame':    189 obs. of  33 variables:
##  $ low                            : int  0 0 0 0 0 0 0 0 0 0 ...
##  $ lowf                           : Factor w/ 2 levels "0","1": 1 1 1 1 1 1 1 1 1 1 ...
##  $ lowfl                          : Factor w/ 2 levels "High","Low": 1 1 1 1 1 1 1 1 1 1 ...
##  $ age                            : int  19 33 20 21 18 21 22 17 29 26 ...
##  $ lwt                            : int  182 155 105 108 107 124 118 103 123 113 ...
##  $ race                           : int  2 3 1 1 1 3 1 3 1 1 ...
##  $ racegr                         : num  1 2 0 0 0 2 0 2 0 0 ...
##  $ racegf                         : Factor w/ 3 levels "0","1","2": 2 3 1 1 1 3 1 3 1 1 ...
##  $ racegfl                        : Factor w/ 3 levels "White","Black",..: 2 3 1 1 1 3 1 3 1 1 ...
##  $ black                          : num  1 0 0 0 0 0 0 0 0 0 ...
##  $ other                          : num  0 1 0 0 0 1 0 1 0 0 ...
##  $ smoke                          : int  0 0 1 1 1 0 0 0 1 1 ...
##  $ smokef                         : Factor w/ 2 levels "0","1": 1 1 2 2 2 1 1 1 2 2 ...
##  $ smokefl                        : Factor w/ 2 levels "Non-smoker","Smoker": 1 1 2 2 2 1 1 1 2 2 ...
##  $ ptl                            : int  0 0 0 0 0 0 0 0 0 0 ...
##  $ ptlgr                          : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ ptlgf                          : Factor w/ 3 levels "0","1","2": 1 1 1 1 1 1 1 1 1 1 ...
##  $ ptlgfl                         : Factor w/ 3 levels "No_preterm_labor",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ one_preterm_labor              : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ twoOrMore_preterm_labor        : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ ht                             : int  0 0 0 0 0 0 0 0 0 0 ...
##  $ htf                            : Factor w/ 2 levels "0","1": 1 1 1 1 1 1 1 1 1 1 ...
##  $ htfl                           : Factor w/ 2 levels "No_hypertension",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ ui                             : int  1 0 0 1 1 0 0 0 0 0 ...
##  $ uif                            : Factor w/ 2 levels "0","1": 2 1 1 2 2 1 1 1 1 1 ...
##  $ uifl                           : Factor w/ 2 levels "No_uterine_irritability",..: 2 1 1 2 2 1 1 1 1 1 ...
##  $ ftv                            : int  0 3 1 2 0 0 1 1 1 0 ...
##  $ ftvgr                          : num  0 2 1 2 0 0 1 1 1 0 ...
##  $ ftvgf                          : Factor w/ 3 levels "0","1","2": 1 3 2 3 1 1 2 2 2 1 ...
##  $ ftvgfl                         : Factor w/ 3 levels "No_physician_visit",..: 1 3 2 3 1 1 2 2 2 1 ...
##  $ one_firstTimester_drVisit      : num  0 0 1 0 0 0 1 1 1 0 ...
##  $ twoOrMore_firstTimester_drVisit: num  0 1 0 1 0 0 0 0 0 0 ...
##  $ bwt                            : int  2523 2551 2557 2594 2600 2622 2637 2637 2663 2665 ...

Variable names

names(bwt.df)

##  [1] "low"                             "lowf"                           
##  [3] "lowfl"                           "age"                            
##  [5] "lwt"                             "race"                           
##  [7] "racegr"                          "racegf"                         
##  [9] "racegfl"                         "black"                          
## [11] "other"                           "smoke"                          
## [13] "smokef"                          "smokefl"                        
## [15] "ptl"                             "ptlgr"                          
## [17] "ptlgf"                           "ptlgfl"                         
## [19] "one_preterm_labor"               "twoOrMore_preterm_labor"        
## [21] "ht"                              "htf"                            
## [23] "htfl"                            "ui"                             
## [25] "uif"                             "uifl"                           
## [27] "ftv"                             "ftvgr"                          
## [29] "ftvgf"                           "ftvgfl"                         
## [31] "one_firstTimester_drVisit"       "twoOrMore_firstTimester_drVisit"
## [33] "bwt"

Total number of variables

sum(table(names(bwt.df))) # Total number of variables: 33

## [1] 33

Make dummies using fastDummies library

library(fastDummies)

# select variables to keep in the data frame and also the variables for which dummies are to be created

# it keeps the variables transformed into dummies

bwt.fd <- 
  
  bwt.df %>% 
  
  dplyr::select(lowfl, age, lwt, racegfl, smokefl, ptlgfl, htfl, uifl, ftvgfl, bwt) %>% 
  
  dummy_cols(
    
    select_columns = c("racegfl", "ptlgfl", "ftvgfl"), 
    
    remove_first_dummy = TRUE
    
    )

str(bwt.fd)

## 'data.frame':    189 obs. of  16 variables:
##  $ lowfl                                 : Factor w/ 2 levels "High","Low": 1 1 1 1 1 1 1 1 1 1 ...
##  $ age                                   : int  19 33 20 21 18 21 22 17 29 26 ...
##  $ lwt                                   : int  182 155 105 108 107 124 118 103 123 113 ...
##  $ racegfl                               : Factor w/ 3 levels "White","Black",..: 2 3 1 1 1 3 1 3 1 1 ...
##  $ smokefl                               : Factor w/ 2 levels "Non-smoker","Smoker": 1 1 2 2 2 1 1 1 2 2 ...
##  $ ptlgfl                                : Factor w/ 3 levels "No_preterm_labor",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ htfl                                  : Factor w/ 2 levels "No_hypertension",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ uifl                                  : Factor w/ 2 levels "No_uterine_irritability",..: 2 1 1 2 2 1 1 1 1 1 ...
##  $ ftvgfl                                : Factor w/ 3 levels "No_physician_visit",..: 1 3 2 3 1 1 2 2 2 1 ...
##  $ bwt                                   : int  2523 2551 2557 2594 2600 2622 2637 2637 2663 2665 ...
##  $ racegfl_Black                         : int  1 0 0 0 0 0 0 0 0 0 ...
##  $ racegfl_Other                         : int  0 1 0 0 0 1 0 1 0 0 ...
##  $ ptlgfl_One_preterm_labor              : int  0 0 0 0 0 0 0 0 0 0 ...
##  $ ptlgfl_TwoOrMore_preterm_labor        : int  0 0 0 0 0 0 0 0 0 0 ...
##  $ ftvgfl_One_firstTimester_drVisit      : int  0 0 1 0 0 0 1 1 1 0 ...
##  $ ftvgfl_twoOrMore_firstTimester_drVisit: int  0 1 0 1 0 0 0 0 0 0 ...

Make dummies using recipes library

# Preprocessing Tools to Create Design Matrices

library(recipes)

## 
## Attaching package: 'recipes'

## The following object is masked from 'package:stringr':
## 
##     fixed

## The following object is masked from 'package:stats':
## 
##     step

# select variables to keep in the data frame and also the variables for which dummies are to be created

# it won't keep the variables transformed into dummies

bwt.dm <- 
  
  bwt.df %>% 
  
  recipe(~ lowfl + age + lwt + racegfl + smokefl + ptlgfl + htfl + uifl + ftvgfl + bwt) %>% 
  
  step_dummy(racegfl, ptlgfl, ftvgfl) %>% 
  
  prep(training = bwt.df) %>% 
  
  bake(new_data = bwt.df)

str(bwt.dm)

## Classes 'tbl_df', 'tbl' and 'data.frame':    189 obs. of  13 variables:
##  $ lowfl                                 : Factor w/ 2 levels "High","Low": 1 1 1 1 1 1 1 1 1 1 ...
##  $ age                                   : int  19 33 20 21 18 21 22 17 29 26 ...
##  $ lwt                                   : int  182 155 105 108 107 124 118 103 123 113 ...
##  $ smokefl                               : Factor w/ 2 levels "Non-smoker","Smoker": 1 1 2 2 2 1 1 1 2 2 ...
##  $ htfl                                  : Factor w/ 2 levels "No_hypertension",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ uifl                                  : Factor w/ 2 levels "No_uterine_irritability",..: 2 1 1 2 2 1 1 1 1 1 ...
##  $ bwt                                   : int  2523 2551 2557 2594 2600 2622 2637 2637 2663 2665 ...
##  $ racegfl_Black                         : num  1 0 0 0 0 0 0 0 0 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ racegfl: chr "contr.treatment"
##  $ racegfl_Other                         : num  0 1 0 0 0 1 0 1 0 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ racegfl: chr "contr.treatment"
##  $ ptlgfl_One_preterm_labor              : num  0 0 0 0 0 0 0 0 0 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ ptlgfl: chr "contr.treatment"
##  $ ptlgfl_TwoOrMore_preterm_labor        : num  0 0 0 0 0 0 0 0 0 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ ptlgfl: chr "contr.treatment"
##  $ ftvgfl_One_firstTimester_drVisit      : num  0 0 1 0 0 0 1 1 1 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ ftvgfl: chr "contr.treatment"
##  $ ftvgfl_twoOrMore_firstTimester_drVisit: num  0 1 0 1 0 0 0 0 0 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ ftvgfl: chr "contr.treatment"

Save the birthwt dataset as a csv file

write_csv(birthwt, "birthwt.csv")

Import the birthwt.csv file as a tibble

bwt_t <- read_csv("birthwt.csv")

## Parsed with column specification:
## cols(
##   low = col_double(),
##   age = col_double(),
##   lwt = col_double(),
##   race = col_double(),
##   smoke = col_double(),
##   ptl = col_double(),
##   ht = col_double(),
##   ui = col_double(),
##   ftv = col_double(),
##   bwt = col_double()
## )

Structure of the imported tibble

str(bwt_t)

## Classes 'spec_tbl_df', 'tbl_df', 'tbl' and 'data.frame': 189 obs. of  10 variables:
##  $ low  : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ age  : num  19 33 20 21 18 21 22 17 29 26 ...
##  $ lwt  : num  182 155 105 108 107 124 118 103 123 113 ...
##  $ race : num  2 3 1 1 1 3 1 3 1 1 ...
##  $ smoke: num  0 0 1 1 1 0 0 0 1 1 ...
##  $ ptl  : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ ht   : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ ui   : num  1 0 0 1 1 0 0 0 0 0 ...
##  $ ftv  : num  0 3 1 2 0 0 1 1 1 0 ...
##  $ bwt  : num  2523 2551 2557 2594 2600 ...
##  - attr(*, "spec")=
##   .. cols(
##   ..   low = col_double(),
##   ..   age = col_double(),
##   ..   lwt = col_double(),
##   ..   race = col_double(),
##   ..   smoke = col_double(),
##   ..   ptl = col_double(),
##   ..   ht = col_double(),
##   ..   ui = col_double(),
##   ..   ftv = col_double(),
##   ..   bwt = col_double()
##   .. )

glimpse(bwt_t)

## Observations: 189
## Variables: 10
## $ low   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...
## $ age   <dbl> 19, 33, 20, 21, 18, 21, 22, 17, 29, 26, 19, 19, 22, 30, ...
## $ lwt   <dbl> 182, 155, 105, 108, 107, 124, 118, 103, 123, 113, 95, 15...
## $ race  <dbl> 2, 3, 1, 1, 1, 3, 1, 3, 1, 1, 3, 3, 3, 3, 1, 1, 2, 1, 3,...
## $ smoke <dbl> 0, 0, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, 1, 0, 1, 0,...
## $ ptl   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,...
## $ ht    <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0,...
## $ ui    <dbl> 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1,...
## $ ftv   <dbl> 0, 3, 1, 2, 0, 0, 1, 1, 1, 0, 0, 1, 0, 2, 0, 0, 0, 3, 0,...
## $ bwt   <dbl> 2523, 2551, 2557, 2594, 2600, 2622, 2637, 2637, 2663, 26...

Transform the tibble

-Factorize

-Recode

-Regroup

-Collapse levels and label the levels

Factorize, recode, collapse the levels of the variables and/or label the levels of the variables

bwt_tcf <- bwt_t %>%
  
  mutate(
    
    lowf=as_factor(low), 
    lowfl=fct_recode(lowf, High="0", Low="1"),
    
    smokef=as_factor(smoke), 
    smokefl=fct_recode(smokef, Non_smoker="0", Smoker="1"),
    
    racef=as_factor(race), 
    racegf=fct_recode(racef, "0"="1", "1"="2", "2"="3"),
    racegfl=fct_recode(racegf, "white"="0",  "black"="1", "other"="2"),
    
    htf=as_factor(ht), 
    htfl=fct_recode(htf, No_hypertension="0", Hypertension="1"),
    
    uif=as_factor(ui), 
    uifl=fct_recode(uif, No_uterine_irritability="0", Uterine_irritability="1"),
    
    ptlf=as_factor(ptl), 
    ptlgf=fct_collapse(ptlf, "0"="0", "1"="1", "2"=c("2", "3")),
    ptlgfl=fct_collapse(ptlf, No_preterm_labor="0", One_preterm_labor="1", Two_plus_preterm_labor=c("2", "3")),
    
    ftvf=as_factor(ftv), 
    ftvgf=fct_collapse(ftvf, "2"=c("2", "3", "4", "6")),
    ftvgfl=fct_collapse(ftvf, No_physician_visit="0", One_physician_visit="1", Two_plus_physician_visit=c("2", "3", "4", "6")),
    
    )

Rearrange the variables the tibble

bwt_tcf <- bwt_tcf %>% dplyr::select(low, lowf, lowfl,
                          age, lwt, 
                          race, racef, racegf,  racegfl,
                          smoke, smokef, smokefl,
                          ptl, ptlf, ptlgf, ptlgfl,
                          ht, htf, htfl,
                          ui, uif, uifl,
                          ftv, ftvf, ftvgf, ftvgfl,
                          bwt
                          )

str(bwt_tcf)

## Classes 'spec_tbl_df', 'tbl_df', 'tbl' and 'data.frame': 189 obs. of  27 variables:
##  $ low    : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ lowf   : Factor w/ 2 levels "0","1": 1 1 1 1 1 1 1 1 1 1 ...
##  $ lowfl  : Factor w/ 2 levels "High","Low": 1 1 1 1 1 1 1 1 1 1 ...
##  $ age    : num  19 33 20 21 18 21 22 17 29 26 ...
##  $ lwt    : num  182 155 105 108 107 124 118 103 123 113 ...
##  $ race   : num  2 3 1 1 1 3 1 3 1 1 ...
##  $ racef  : Factor w/ 3 levels "1","2","3": 2 3 1 1 1 3 1 3 1 1 ...
##  $ racegf : Factor w/ 3 levels "0","1","2": 2 3 1 1 1 3 1 3 1 1 ...
##  $ racegfl: Factor w/ 3 levels "white","black",..: 2 3 1 1 1 3 1 3 1 1 ...
##  $ smoke  : num  0 0 1 1 1 0 0 0 1 1 ...
##  $ smokef : Factor w/ 2 levels "0","1": 1 1 2 2 2 1 1 1 2 2 ...
##  $ smokefl: Factor w/ 2 levels "Non_smoker","Smoker": 1 1 2 2 2 1 1 1 2 2 ...
##  $ ptl    : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ ptlf   : Factor w/ 4 levels "0","1","2","3": 1 1 1 1 1 1 1 1 1 1 ...
##  $ ptlgf  : Factor w/ 3 levels "0","1","2": 1 1 1 1 1 1 1 1 1 1 ...
##  $ ptlgfl : Factor w/ 3 levels "No_preterm_labor",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ ht     : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ htf    : Factor w/ 2 levels "0","1": 1 1 1 1 1 1 1 1 1 1 ...
##  $ htfl   : Factor w/ 2 levels "No_hypertension",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ ui     : num  1 0 0 1 1 0 0 0 0 0 ...
##  $ uif    : Factor w/ 2 levels "0","1": 2 1 1 2 2 1 1 1 1 1 ...
##  $ uifl   : Factor w/ 2 levels "No_uterine_irritability",..: 2 1 1 2 2 1 1 1 1 1 ...
##  $ ftv    : num  0 3 1 2 0 0 1 1 1 0 ...
##  $ ftvf   : Factor w/ 6 levels "0","1","2","3",..: 1 4 2 3 1 1 2 2 2 1 ...
##  $ ftvgf  : Factor w/ 3 levels "0","1","2": 1 3 2 3 1 1 2 2 2 1 ...
##  $ ftvgfl : Factor w/ 3 levels "No_physician_visit",..: 1 3 2 3 1 1 2 2 2 1 ...
##  $ bwt    : num  2523 2551 2557 2594 2600 ...

names(bwt_tcf)

##  [1] "low"     "lowf"    "lowfl"   "age"     "lwt"     "race"    "racef"  
##  [8] "racegf"  "racegfl" "smoke"   "smokef"  "smokefl" "ptl"     "ptlf"   
## [15] "ptlgf"   "ptlgfl"  "ht"      "htf"     "htfl"    "ui"      "uif"    
## [22] "uifl"    "ftv"     "ftvf"    "ftvgf"   "ftvgfl"  "bwt"

Make dummies using fastDummies libray

# outputs all variables plus dummy variables

bwt_fd <- bwt_tcf %>% dummy_cols(
  
  select_columns = c("racegfl", "ptlgfl", "ftvgfl"), 
  
  remove_first_dummy = TRUE)

str(bwt_fd)

## Classes 'tbl_df', 'tbl' and 'data.frame':    189 obs. of  33 variables:
##  $ low                            : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ lowf                           : Factor w/ 2 levels "0","1": 1 1 1 1 1 1 1 1 1 1 ...
##  $ lowfl                          : Factor w/ 2 levels "High","Low": 1 1 1 1 1 1 1 1 1 1 ...
##  $ age                            : num  19 33 20 21 18 21 22 17 29 26 ...
##  $ lwt                            : num  182 155 105 108 107 124 118 103 123 113 ...
##  $ race                           : num  2 3 1 1 1 3 1 3 1 1 ...
##  $ racef                          : Factor w/ 3 levels "1","2","3": 2 3 1 1 1 3 1 3 1 1 ...
##  $ racegf                         : Factor w/ 3 levels "0","1","2": 2 3 1 1 1 3 1 3 1 1 ...
##  $ racegfl                        : Factor w/ 3 levels "white","black",..: 2 3 1 1 1 3 1 3 1 1 ...
##  $ smoke                          : num  0 0 1 1 1 0 0 0 1 1 ...
##  $ smokef                         : Factor w/ 2 levels "0","1": 1 1 2 2 2 1 1 1 2 2 ...
##  $ smokefl                        : Factor w/ 2 levels "Non_smoker","Smoker": 1 1 2 2 2 1 1 1 2 2 ...
##  $ ptl                            : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ ptlf                           : Factor w/ 4 levels "0","1","2","3": 1 1 1 1 1 1 1 1 1 1 ...
##  $ ptlgf                          : Factor w/ 3 levels "0","1","2": 1 1 1 1 1 1 1 1 1 1 ...
##  $ ptlgfl                         : Factor w/ 3 levels "No_preterm_labor",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ ht                             : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ htf                            : Factor w/ 2 levels "0","1": 1 1 1 1 1 1 1 1 1 1 ...
##  $ htfl                           : Factor w/ 2 levels "No_hypertension",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ ui                             : num  1 0 0 1 1 0 0 0 0 0 ...
##  $ uif                            : Factor w/ 2 levels "0","1": 2 1 1 2 2 1 1 1 1 1 ...
##  $ uifl                           : Factor w/ 2 levels "No_uterine_irritability",..: 2 1 1 2 2 1 1 1 1 1 ...
##  $ ftv                            : num  0 3 1 2 0 0 1 1 1 0 ...
##  $ ftvf                           : Factor w/ 6 levels "0","1","2","3",..: 1 4 2 3 1 1 2 2 2 1 ...
##  $ ftvgf                          : Factor w/ 3 levels "0","1","2": 1 3 2 3 1 1 2 2 2 1 ...
##  $ ftvgfl                         : Factor w/ 3 levels "No_physician_visit",..: 1 3 2 3 1 1 2 2 2 1 ...
##  $ bwt                            : num  2523 2551 2557 2594 2600 ...
##  $ racegfl_black                  : int  1 0 0 0 0 0 0 0 0 0 ...
##  $ racegfl_other                  : int  0 1 0 0 0 1 0 1 0 0 ...
##  $ ptlgfl_One_preterm_labor       : int  0 0 0 0 0 0 0 0 0 0 ...
##  $ ptlgfl_Two_plus_preterm_labor  : int  0 0 0 0 0 0 0 0 0 0 ...
##  $ ftvgfl_One_physician_visit     : int  0 0 1 0 0 0 1 1 1 0 ...
##  $ ftvgfl_Two_plus_physician_visit: int  0 1 0 1 0 0 0 0 0 0 ...
##  - attr(*, ".internal.selfref")=<externalptr>

# Total number of variables: 45

sum(table(names(bwt_fd))) 

## [1] 33

Make dummies using recipes libray

# recipe selects some of the variables in the tibble

# won't keep the original variables transformed into dummies

bwt_dm <- 
  
  bwt_tcf %>% 
  
  recipe(~ lowfl + age + lwt + racegfl + smokefl + ptlgfl + htfl + uifl + ftvgfl + bwt) %>% 
  
  step_dummy(racegfl, ptlgfl, ftvgfl) %>% 
  
  prep(training = bwt_tcf) %>% 
  
  bake(new_data = bwt_tcf)

str(bwt_dm)

## Classes 'tbl_df', 'tbl' and 'data.frame':    189 obs. of  13 variables:
##  $ lowfl                          : Factor w/ 2 levels "High","Low": 1 1 1 1 1 1 1 1 1 1 ...
##  $ age                            : num  19 33 20 21 18 21 22 17 29 26 ...
##  $ lwt                            : num  182 155 105 108 107 124 118 103 123 113 ...
##  $ smokefl                        : Factor w/ 2 levels "Non_smoker","Smoker": 1 1 2 2 2 1 1 1 2 2 ...
##  $ htfl                           : Factor w/ 2 levels "No_hypertension",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ uifl                           : Factor w/ 2 levels "No_uterine_irritability",..: 2 1 1 2 2 1 1 1 1 1 ...
##  $ bwt                            : num  2523 2551 2557 2594 2600 ...
##  $ racegfl_black                  : num  1 0 0 0 0 0 0 0 0 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ racegfl: chr "contr.treatment"
##  $ racegfl_other                  : num  0 1 0 0 0 1 0 1 0 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ racegfl: chr "contr.treatment"
##  $ ptlgfl_One_preterm_labor       : num  0 0 0 0 0 0 0 0 0 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ ptlgfl: chr "contr.treatment"
##  $ ptlgfl_Two_plus_preterm_labor  : num  0 0 0 0 0 0 0 0 0 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ ptlgfl: chr "contr.treatment"
##  $ ftvgfl_One_physician_visit     : num  0 0 1 0 0 0 1 1 1 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ ftvgfl: chr "contr.treatment"
##  $ ftvgfl_Two_plus_physician_visit: num  0 1 0 1 0 0 0 0 0 0 ...
##   ..- attr(*, "assign")= int  0 1 1
##   ..- attr(*, "contrasts")=List of 1
##   .. ..$ ftvgfl: chr "contr.treatment"

Sample observations

head(birthwt, n=3)

	low	age	lwt	race	smoke	ptl	ht	ui	ftv	bwt
85	0	19	182	2	0	0	0	1	0	2523
86	0	33	155	3	0	0	0	0	3	2551
87	0	20	105	1	1	0	0	0	1	2557

head(bwt.df, n=3)

	low	lowf	lowfl	age	lwt	race	racegr	racegf	racegfl	black	other	smoke	smokef	smokefl	ptl	ptlgr	ptlgf	ptlgfl	one_preterm_labor	twoOrMore_preterm_labor	ht	htf	htfl	ui	uif	uifl	ftv	ftvgr	ftvgf	ftvgfl	one_firstTimester_drVisit	twoOrMore_firstTimester_drVisit	bwt
85	0	0	High	19	182	2	1	1	Black	1	0	0	0	Non-smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit	0	0	2523
86	0	0	High	33	155	3	2	2	Other	0	1	0	0	Non-smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	0	0	No_uterine_irritability	3	2	2	twoOrMore_firstTimester_drVisit	0	1	2551
87	0	0	High	20	105	1	0	0	White	0	0	1	1	Smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	0	0	No_uterine_irritability	1	1	1	One_firstTimester_drVisit	1	0	2557

birthwt %>% sample_n(2, replace = TRUE)

low	age	lwt	race	smoke	ptl	ht	ui	ftv	bwt
1	25	92	1	1	0	0	0	0	1928
1	20	150	1	1	0	0	0	2	1928

# glimpse the 1st 3 rows of the tibble
# Note there are no row or observation numbers.

head(bwt_t, n=3)

low	age	lwt	race	smoke	ptl	ht	ui	ftv	bwt
0	19	182	2	0	0	0	1	0	2523
0	33	155	3	0	0	0	0	3	2551
0	20	105	1	1	0	0	0	1	2557

# sample two observations

bwt_t %>% sample_n(2)

low	age	lwt	race	smoke	ptl	ht	ui	ftv	bwt
0	28	130	3	0	0	0	0	0	3969
0	29	150	1	0	0	0	0	2	2920

# first 3 rows

bwt_tcf[1:3,]

low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl	bwt
0	0	High	19	182	2	2	1	black	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit	2523
0	0	High	33	155	3	3	2	other	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	3	3	2	Two_plus_physician_visit	2551
0	0	High	20	105	1	1	0	white	1	1	Smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	1	1	1	One_physician_visit	2557

# sample 1% of the tibble

bwt_tcf %>% sample_frac(.01)

low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl	bwt
1	1	Low	21	100	3	3	2	other	0	0	Non_smoker	1	1	1	One_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	4	4	2	Two_plus_physician_visit	2301
0	0	High	35	170	1	1	0	white	0	0	Non_smoker	1	1	1	One_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	1	1	1	One_physician_visit	4174

# rows 1, 10 and 100

bwt_tcf[c(1, 10, 100),]

low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl	bwt
0	0	High	19	182	2	2	1	black	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit	2523
0	0	High	26	113	1	1	0	white	1	1	Smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	0	0	0	No_physician_visit	2665
0	0	High	30	137	1	1	0	white	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	1	1	1	One_physician_visit	3699

Sample variables

# output: a vector of length 189

bwt_tcf[["bwt"]] %>% length()          

## [1] 189

# output: a vector of length 189

bwt_tcf$bwt %>% length()       

## [1] 189

# output: a vector of 3 obsrvations from variable "bwt"

bwt_tcf$bwt[c(1, 3, 6)]    

## [1] 2523 2557 2622

Select variables of your choice

## select race, smoke and bwt variables

birthwt[c(4, 5, 10)] %>% head(n=2)

	race	smoke	bwt
85	2	0	2523
86	3	0	2551

bwt.df[c(6, 12, 33)] %>% head(n=2)

	race	smoke	bwt
85	2	0	2523
86	3	0	2551

## select race, smoke and bwt variables

bwt_tcf %>% dplyr::select(race, smoke, bwt) %>% head(n=2)

race	smoke	bwt
2	0	2523
3	0	2551

Select continous variables

bwt_tcf %>% dplyr::select(age, lwt, bwt) %>% head(n=2)

age	lwt	bwt
19	182	2523
33	155	2551

Select variables that ends_with “t”

bwt_tcf %>% dplyr::select(ends_with("t")) %>% head(n=2)

lwt	ht	bwt
182	0	2523
155	0	2551

Drop few variables

bwt_tcf %>% dplyr::select(-c(low, race, smoke, ptl, ht, ui, ftv)) %>% head(n=2)

lowf	lowfl	age	lwt	racef	racegf	racegfl	smokef	smokefl	ptlf	ptlgf	ptlgfl	htf	htfl	uif	uifl	ftvf	ftvgf	ftvgfl	bwt
0	High	19	182	2	1	black	0	Non_smoker	0	0	No_preterm_labor	0	No_hypertension	1	Uterine_irritability	0	0	No_physician_visit	2523
0	High	33	155	3	2	other	0	Non_smoker	0	0	No_preterm_labor	0	No_hypertension	0	No_uterine_irritability	3	2	Two_plus_physician_visit	2551

Obtain or filter or subset rows/observations from a dataset based on certain conditions

# output: a vector of 96 obsrvations for variable "racegfl" where the level is "white"

bwt_tcf$bwt[bwt_tcf$racegfl=="white"] %>% length()   

## [1] 96

# output: "bwt" as a list, data.frame

bwt_tcf["bwt"] %>% dim()                

## [1] 189   1

# output 3 observations/rows for the "bwt" variables

bwt_tcf[c(1,3,6), "bwt"]             

bwt
2523
2557
2622

# output: a data.frame when race=="White"

bwt_tcf[bwt_tcf$racegfl=="white", ]   %>% dim()   

## [1] 96 27

Identify the observation or row number for a particular value (or values of a ) variable

which(bwt.df$bwt==2557)

## [1] 3

bwt.df[bwt.df$bwt==2557,]

	low	lowf	lowfl	age	lwt	race	racegr	racegf	racegfl	black	other	smoke	smokef	smokefl	ptl	ptlgr	ptlgf	ptlgfl	one_preterm_labor	twoOrMore_preterm_labor	ht	htf	htfl	ui	uif	uifl	ftv	ftvgr	ftvgf	ftvgfl	one_firstTimester_drVisit	twoOrMore_firstTimester_drVisit	bwt
87	0	0	High	20	105	1	0	0	White	0	0	1	1	Smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	0	0	No_uterine_irritability	1	1	1	One_firstTimester_drVisit	1	0	2557

which(bwt_tcf$bwt==2557)

## [1] 3

bwt_tcf[bwt_tcf$bwt==2557,]

low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl	bwt
0	0	High	20	105	1	1	0	white	1	1	Smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	1	1	1	One_physician_visit	2557

Identify observations where a variable has missing value

bwt_tcf[is.na(bwt_tcf$age),] %>% count()

n
0

filter(bwt_tcf, is.na(age)) %>% count()

n
0

bwt_tcf %>% filter(is.na(age)) %>% count()

n
0

Identify observations for non-missing values of a variable

bwt_tcf[!is.na(bwt_tcf$age), ] %>% count()

n
189

bwt_tcf %>% filter(!is.na(age)) %>% count()

n
189

sum(!is.na(bwt_tcf$age))

## [1] 189

Identify observations where the values of a variable are compared to certain statistical measures

-Suitable for continous variables. For eaxmple, when age is less than or equal to the mean age.

# observations where age is less than or equal to the mean age 


bwt.df[bwt.df$age <= mean(bwt.df$age), ] %>% count()

n
107

bwt_tcf[bwt_tcf$age <= mean(bwt_tcf$age), ] %>% count()

n
107

bwt_tcf %>% filter(age<=mean(age)) %>% count()

n
107

# observations where age is greater than the mean age

bwt_tcf %>% filter(age > mean(age)) %>%  count() 

n
82

# observations where age is between 30 and 40 (must use "&")


bwt.df %>% subset(age >= 30 & age <= 40) %>% count()

n
26

# observations where age is between 30 and 40 for a categorical variable


bwt_tcf %>% filter(between(age, 30, 40)) %>% count(lowfl)

lowfl	n
High	22
Low	4

bwt_tcf %>% filter(age >= 30, age <= 40) %>% count(racegfl)

racegfl	n
white	18
black	2
other	6

Identify observations for various levels of a categorical variable

bwt.df[bwt.df$race==2, ] %>% count(race)

race	n
2	26

bwt.df[bwt.df$racegr==1, ] %>% count(race)

race	n
2	26

bwt.df[bwt.df$racegf==1, ] %>% count(race)

race	n
2	26

bwt.df[bwt.df$racegfl=="Black", ] %>% count(race)

race	n
2	26

bwt.df %>% subset(racegfl=="Black") %>% count(race)

race	n
2	26

bwt_tcf[bwt_tcf$racegfl=="white", ] %>% count(race)

race	n
1	96

bwt_tcf %>% subset(racegfl=="white") %>% count(race)

race	n
1	96

bwt_tcf %>% filter(racegfl=="white") %>% count(race)

race	n
1	96

Identify observations that meets more than one subsetting conditions

# observations where race is white or black


bwt_tcf %>% filter(racegfl=="white" | racegfl=="black") %>% count() # 122

n
122

bwt_tcf %>% filter(racegfl=="white" | racegfl=="black") %>% count(race) # 96+26

race	n
1	96
2	26

bwt_tcf %>% filter(racegfl %in% c("white", "black")) %>% count() #122

n
122

bwt_tcf %>% filter(racegfl %in% c("white", "black")) %>% count(race) #96+26

race	n
1	96
2	26

# observations where race is neither white nor black


bwt_tcf[bwt_tcf$racegfl!="white" & bwt_tcf$racegfl!="black",] %>% count() # 67

n
67

bwt_tcf %>% subset(!(racegfl %in% c("white", "black")) ) %>%  count()#67

n
67

bwt_tcf %>% filter(racegfl!="white" & racegfl!="black") %>% count()  #67

n
67

bwt_tcf %>% filter(!(racegfl %in% c("white", "black"))) %>% count()  #67

n
67

# observations where black folks also smoke


bwt_tcf %>% filter(racegfl=="black", smokefl=="Smoker") %>% count(race, smoke) #10

race	smoke	n
2	1	10

# observations where race is not either 1 or 2

bwt_tcf[bwt_tcf$race!=1 & bwt_tcf$race!=2,] %>% count() # 67

n
67

subset(bwt_tcf, !(race %in% c(1, 2)) ) %>%  count() # 67

n
67

bwt_tcf %>% filter(race!=1 & race!=2) %>% count()  #67

n
67

bwt_tcf %>% filter(!(race==1 | race==2)) %>% count()  #67

n
67

bwt_tcf %>% filter(!(race %in% c(1, 2))) %>% count()  #67

n
67

Arrange (sort descending/ascending) variables of interest

-NA are always sorted to the end for local data

-Default is ascending

-desc denotes descending

names(bwt.df)

##  [1] "low"                             "lowf"                           
##  [3] "lowfl"                           "age"                            
##  [5] "lwt"                             "race"                           
##  [7] "racegr"                          "racegf"                         
##  [9] "racegfl"                         "black"                          
## [11] "other"                           "smoke"                          
## [13] "smokef"                          "smokefl"                        
## [15] "ptl"                             "ptlgr"                          
## [17] "ptlgf"                           "ptlgfl"                         
## [19] "one_preterm_labor"               "twoOrMore_preterm_labor"        
## [21] "ht"                              "htf"                            
## [23] "htfl"                            "ui"                             
## [25] "uif"                             "uifl"                           
## [27] "ftv"                             "ftvgr"                          
## [29] "ftvgf"                           "ftvgfl"                         
## [31] "one_firstTimester_drVisit"       "twoOrMore_firstTimester_drVisit"
## [33] "bwt"

# sort by decreasing age

bwt.df[order(bwt.df$age, decreasing = TRUE), ] %>% head(n=5)

	low	lowf	lowfl	age	lwt	race	racegr	racegf	racegfl	black	other	smoke	smokef	smokefl	ptl	ptlgr	ptlgf	ptlgfl	one_preterm_labor	twoOrMore_preterm_labor	ht	htf	htfl	ui	uif	uifl	ftv	ftvgr	ftvgf	ftvgfl	one_firstTimester_drVisit	twoOrMore_firstTimester_drVisit	bwt
226	0	0	High	45	123	1	0	0	White	0	0	0	0	Non-smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	0	0	No_uterine_irritability	1	1	1	One_firstTimester_drVisit	1	0	4990
108	0	0	High	36	202	1	0	0	White	0	0	0	0	Non-smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	0	0	No_uterine_irritability	1	1	1	One_firstTimester_drVisit	1	0	2836
183	0	0	High	36	175	1	0	0	White	0	0	0	0	Non-smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	0	0	No_uterine_irritability	0	0	0	No_physician_visit	0	0	3600
119	0	0	High	35	121	2	1	1	Black	1	0	1	1	Smoker	1	1	1	One_preterm_labor	1	0	0	0	No_hypertension	0	0	No_uterine_irritability	1	1	1	One_firstTimester_drVisit	1	0	2948
223	0	0	High	35	170	1	0	0	White	0	0	0	0	Non-smoker	1	1	1	One_preterm_labor	1	0	0	0	No_hypertension	0	0	No_uterine_irritability	1	1	1	One_firstTimester_drVisit	1	0	4174

# sort by decreasing age and show only age and lwt

bwt.df[order(bwt.df$age, decreasing = TRUE), ][c(4, 5)] %>% head(n=5)  # show columns or variables 10 $ 33

	age	lwt
226	45	123
108	36	202
183	36	175
119	35	121
223	35	170

# sort by age, lwt and bwt


bwt_tcf[order(bwt_tcf$age, bwt_tcf$lwt, bwt_tcf$bwt), ][c(4, 5, 27)] %>% head(n=5)

age	lwt	bwt
14	100	2495
14	101	2466
14	135	3941
15	98	2778
15	110	2353

# sort by  bwt

bwt_tcf %>% dplyr::select(bwt) %>% arrange(bwt) %>% head(n=5)

bwt
709
1021
1135
1330
1474

bwt_tcf %>% dplyr::select(bwt) %>% arrange(desc(bwt)) %>% head(n=5)

bwt
4990
4593
4238
4174
4167

# sort by age, lwt and bwt


bwt_tcf %>% dplyr::select(age, lwt, bwt) %>% arrange(age, lwt, bwt) %>% head(n=5)

age	lwt	bwt
14	100	2495
14	101	2466
14	135	3941
15	98	2778
15	110	2353

bwt_tcf %>% dplyr::select(age, lwt, bwt) %>% arrange(desc(age), desc(lwt), desc(bwt)) %>% head(n=5)

age	lwt	bwt
45	123	4990
36	202	2836
36	175	3600
35	170	4174
35	121	2948

Reorder the position of the variables

bwt.df %>% dplyr::select(bwt, low, lowf, lowfl, everything()) %>% head(n=2)

	bwt	low	lowf	lowfl	age	lwt	race	racegr	racegf	racegfl	black	other	smoke	smokef	smokefl	ptl	ptlgr	ptlgf	ptlgfl	one_preterm_labor	twoOrMore_preterm_labor	ht	htf	htfl	ui	uif	uifl	ftv	ftvgr	ftvgf	ftvgfl	one_firstTimester_drVisit	twoOrMore_firstTimester_drVisit
85	2523	0	0	High	19	182	2	1	1	Black	1	0	0	0	Non-smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit	0	0
86	2551	0	0	High	33	155	3	2	2	Other	0	1	0	0	Non-smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	0	0	No_uterine_irritability	3	2	2	twoOrMore_firstTimester_drVisit	0	1

bwt_tcf %>% dplyr::select(bwt, low, lowf, lowfl, everything()) %>% head(n=2)

bwt	low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl
2523	0	0	High	19	182	2	2	1	black	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit
2551	0	0	High	33	155	3	3	2	other	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	3	3	2	Two_plus_physician_visit

Rename a variable

bwt_tcf %>% rename(body_weight=bwt) %>% dplyr::select(body_weight, everything()) %>% head(n=2)

body_weight	low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl
2523	0	0	High	19	182	2	2	1	black	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit
2551	0	0	High	33	155	3	3	2	other	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	3	3	2	Two_plus_physician_visit

Mutate

quantile(bwt_tcf$bwt)

##   0%  25%  50%  75% 100% 
##  709 2414 2977 3487 4990

range(bwt_tcf$bwt)

## [1]  709 4990

Dichotomize or categorize a continous variable using if_else function

# outcome is a "numeric" variable

bwt_tcf %>% 
  mutate(bwt_category =if_else(bwt <= 2500, 1, 0)) %>%
  dplyr::select(bwt, bwt_category, everything()) %>% 
  head(n=2)

bwt	bwt_category	low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl
2523	0	0	0	High	19	182	2	2	1	black	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit
2551	0	0	0	High	33	155	3	3	2	other	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	3	3	2	Two_plus_physician_visit

# outcome is  numeral "factor" variable

bwt_tcf %>% 
  mutate(bwt_category =as_factor(if_else(bwt <= 2500, 1, 0))) %>%
  dplyr::select(bwt, bwt_category, everything()) %>% 
  head(n=2)

bwt	bwt_category	low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl
2523	0	0	0	High	19	182	2	2	1	black	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit
2551	0	0	0	High	33	155	3	3	2	other	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	3	3	2	Two_plus_physician_visit

# outcome is a "character" variable

bwt_tcf %>% 
  mutate(bwt_category = if_else(bwt <= 2500, "low", "high")) %>%
  dplyr::select(bwt, bwt_category, everything()) %>% 
  head(n=2)

bwt	bwt_category	low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl
2523	high	0	0	High	19	182	2	2	1	black	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit
2551	high	0	0	High	33	155	3	3	2	other	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	3	3	2	Two_plus_physician_visit

# outcome is a character "factor" variable

bwt_tcf %>% 
  mutate(bwt_category = as_factor (if_else(bwt <= 2500, "low", "high"))) %>%
  dplyr::select(bwt, bwt_category, everything()) %>% 
  head(n=2)

bwt	bwt_category	low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl
2523	high	0	0	High	19	182	2	2	1	black	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit
2551	high	0	0	High	33	155	3	3	2	other	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	3	3	2	Two_plus_physician_visit

Dichotomize a continuous variable using cut function

# dichotomize bwt 

bwt.df$bwt_category <- cut(bwt.df$bwt, breaks=c(-Inf, 2500, Inf), labels=c("low","high"))

bwt.df %>% count(bwt_category)

bwt_category	n
low	59
high	130

table(bwt.df$lowfl, bwt.df$bwt_category)

##       
##        low high
##   High   0  130
##   Low   59    0

boxplot(bwt.df$bwt ~ bwt.df$bwt_category, col=2:3, horizontal = TRUE)

bwt_tcf <- bwt_tcf %>% mutate(bwt_category=cut(bwt, breaks=c(-Inf, 2500, Inf), labels=c("low","high")))

boxplot(bwt_tcf$bwt ~ bwt_tcf$bwt_category, col=2:3, horizontal = TRUE)

Multi-categorize a contnous variable using cut function

bwt.df$bwt %>% quantile()

##   0%  25%  50%  75% 100% 
##  709 2414 2977 3487 4990

# categorize using `cut` function into 4 arbitrary levels

bwt.df$bwt_cut <- cut(bwt.df$bwt, 4, labels=c("wt1", "wt2", "wt3", "wt4"))


bwt.df %>% count(bwt_cut)

bwt_cut	n
wt1	9
wt2	73
wt3	91
wt4	16

# Cut into 4 groups (using quantile values)

bwt.df$bwt_category <- cut(bwt.df$bwt, 
                   breaks=c(-Inf, 2414, 2977, 3487, Inf), 
                   labels=c("wt1", "wt2", "wt3", "wt4"))


bwt.df %>% count(bwt_category)

bwt_category	n
wt1	48
wt2	49
wt3	45
wt4	47

boxplot(bwt.df$bwt ~ bwt.df$bwt_category, col=5:8, horizontal = TRUE)

# use quantile probablilities instead of calculated values to cut into 4 groups

bwt.df$bwt_quantiles <- cut(bwt.df$bwt, 
                   breaks=quantile(bwt.df$bwt, probs = seq(0, 1, 0.25)), 
                   include.lowest = TRUE,
                   labels=c(1:4))

bwt.df %>% count(bwt_quantiles)

bwt_quantiles	n
1	48
2	49
3	45
4	47

bwt.df$bwt_quantiles <- cut(bwt.df$bwt, 
                   breaks=quantile(bwt.df$bwt, probs = c(0, 0.25, 0.5, 0.75, 1)), 
                   include.lowest = TRUE,
                   labels=c("wt1", "wt2", "wt3", "wt4"))


bwt.df %>% count(bwt_quantiles)

bwt_quantiles	n
wt1	48
wt2	49
wt3	45
wt4	47

Create ordered factors from a continous variable

bwt.df <- bwt.df %>% mutate(bwt_tertile=cut(bwt, 
                                  breaks=quantile(bwt, probs = c(0, 0.337, 0.663, 1)), 
                                  include.lowest = TRUE,
                                  labels=c(3:1)))

bwt.df %>% count(bwt_tertile)

bwt_tertile	n
3	64
2	62
1	63

boxplot(bwt.df$bwt ~ bwt.df$bwt_tertile, col=5:7, horizontal = TRUE )

bwt.df <- bwt.df %>% 
  mutate(
  #bwt_tertile_recode=fct_recode(bwt_tertile, Low="3", Medium="2", High="1"),
  #bwt_tertile_order=factor(bwt_tertile_recode, order=TRUE),
  #bwt_tertile_order=ordered(bwt_tertile_recode)
    bwt_tertile_order=factor(bwt_tertile, levels=c("1", "2", "3"), labels =c("low", "medium", "high"), order = TRUE )
  
  )

bwt.df %>% count(bwt_tertile_order)

bwt_tertile_order	n
low	63
medium	62
high	64

boxplot(bwt.df$bwt ~ bwt.df$bwt_tertile_order, col=5:7, horizontal = TRUE )

Data transformation

# create numeric, logical and factor variable

bwt_tcf %>% 
  mutate(bwt_in_kg=bwt/1000, 
         low_bwt=bwt_in_kg <= 2.5,
         low_wt_class = as_factor(if_else(low_bwt, 1, 0))) %>% 
  dplyr::select(bwt, bwt_in_kg, low_bwt, low_wt_class) %>% 
  head(n=2)

bwt	bwt_in_kg	low_bwt	low_wt_class
2523	2.523	FALSE	0
2551	2.551	FALSE	0

Introduce NA

bwt_tcf %>% 
  mutate(bwt_value = ifelse(bwt <= 2500, NA, bwt)) %>%
  arrange(bwt_value) %>% 
  dplyr::select(bwt, bwt_value, everything()) %>% tail(n=2)

bwt	bwt_value	low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl	bwt_category
2495	NA	1	1	Low	17	142	2	2	1	black	0	0	Non_smoker	0	0	0	No_preterm_labor	1	1	Hypertension	0	0	No_uterine_irritability	0	0	0	No_physician_visit	low
2495	NA	1	1	Low	21	130	1	1	0	white	1	1	Smoker	0	0	0	No_preterm_labor	1	1	Hypertension	0	0	No_uterine_irritability	3	3	2	Two_plus_physician_visit	low

bwt_tcf %>% 
  mutate(bwt_new = ifelse(bwt < 2600 | bwt > 4000, NA, bwt)) %>% 
  dplyr::select(bwt, bwt_new, everything()) %>% 
  head(n=2)

bwt	bwt_new	low	lowf	lowfl	age	lwt	race	racef	racegf	racegfl	smoke	smokef	smokefl	ptl	ptlf	ptlgf	ptlgfl	ht	htf	htfl	ui	uif	uifl	ftv	ftvf	ftvgf	ftvgfl	bwt_category
2523	NA	0	0	High	19	182	2	2	1	black	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit	high
2551	NA	0	0	High	33	155	3	3	2	other	0	0	Non_smoker	0	0	0	No_preterm_labor	0	0	No_hypertension	0	0	No_uterine_irritability	3	3	2	Two_plus_physician_visit	high

Transmute

# keep only the new variables


bwt_tcf %>% 
  transmute(bwt_kg=bwt/1000, 
                   low_wt=bwt_kg <=2.5,
                   low_wt_class = as_factor(if_else(low_wt, 1, 0))) %>% 
  head(n=2)

bwt_kg	low_wt	low_wt_class
2.523	FALSE	0
2.551	FALSE	0

Summarise

Categorical variables: counts and frequencies

# output "tbl_df"     "tbl"        "data.frame"

bwt_tcf %>% count(race)

race	n
1	96
2	26
3	67

bwt_tcf %>% count(low, race)

low	race	n
0	1	73
0	2	15
0	3	42
1	1	23
1	2	11
1	3	25

# output table

table(bwt_tcf$race)

## 
##  1  2  3 
## 96 26 67

table(bwt_tcf$low, bwt_tcf$race)

##    
##      1  2  3
##   0 73 15 42
##   1 23 11 25

Frequencies of the factor variables in the tibble

table(bwt_tcf$ptl)

## 
##   0   1   2   3 
## 159  24   5   1

table(bwt_tcf$ptlf)

## 
##   0   1   2   3 
## 159  24   5   1

table(bwt_tcf$ptlgf)

## 
##   0   1   2 
## 159  24   6

table(bwt_tcf$ptlgfl)

## 
##       No_preterm_labor      One_preterm_labor Two_plus_preterm_labor 
##                    159                     24                      6

table(bwt_tcf$ftv)

## 
##   0   1   2   3   4   6 
## 100  47  30   7   4   1

table(bwt_tcf$ftvf)

## 
##   0   1   2   3   4   6 
## 100  47  30   7   4   1

table(bwt_tcf$ftvgf)

## 
##   0   1   2 
## 100  47  42

table(bwt_tcf$ftvgfl)

## 
##       No_physician_visit      One_physician_visit Two_plus_physician_visit 
##                      100                       47                       42

table(bwt_tcf$race)

## 
##  1  2  3 
## 96 26 67

table(bwt_tcf$racef)

## 
##  1  2  3 
## 96 26 67

table(bwt_tcf$racegf)

## 
##  0  1  2 
## 96 26 67

table(bwt_tcf$racegfl)

## 
## white black other 
##    96    26    67

Continous variables (without any grouping)

# individual stats

nrow(table(bwt_tcf$bwt)) 

## [1] 131

min(bwt_tcf$bwt)

## [1] 709

max(bwt_tcf$bwt)       

## [1] 4990

range(bwt_tcf$bwt)             

## [1]  709 4990

mean(bwt_tcf$bwt)

## [1] 2944.587

median(bwt_tcf$bwt)

## [1] 2977

sd(bwt_tcf$bwt)

## [1] 729.2143

mad(bwt_tcf$bwt)                  # median absolute deviation

## [1] 834.7038

var(bwt_tcf$bwt)

## [1] 531753.5

IQR(bwt_tcf$bwt)                  # interquartile range

## [1] 1073

quantile(bwt_tcf$bwt)             # quartiles (by default)

##   0%  25%  50%  75% 100% 
##  709 2414 2977 3487 4990

quantile(bwt_tcf$bwt, c(1, 3)/4)  # specific percentiles (25% & 75% in this case)

##  25%  75% 
## 2414 3487

kurtosi(bwt_tcf$bwt)             # psych package

## [1] -0.1436834

skew(bwt_tcf$bwt)                # psych package

## [1] -0.205337

# use summarise function

bwt_tcf %>% 
  summarise(no_of_total_obs=n(),
            no_of_distinct_bwt_obs=n_distinct(bwt),
            
            bwt_min=min(bwt, na.rm = TRUE),
            bwt_max=max(bwt, na.rm = TRUE),
            bwt_50th_quantile=quantile(bwt, 0.5, na.rm = TRUE),
            
            bwt_avg=mean(bwt, na.rm = TRUE), 
            bwt_median=median(bwt, na.rm = TRUE),
            
            bwt_sd=sd(bwt, na.rm = TRUE),
            bwt_IQR=IQR(bwt, na.rm = TRUE),
            bwt_mad=mad(bwt, na.rm = TRUE),
            
            bwt_tirst=first(bwt),
            bwt_last=last(bwt),
            bwt_4th=nth(bwt, 4)) %>%
  t

##                             [,1]
## no_of_total_obs         189.0000
## no_of_distinct_bwt_obs  131.0000
## bwt_min                 709.0000
## bwt_max                4990.0000
## bwt_50th_quantile      2977.0000
## bwt_avg                2944.5873
## bwt_median             2977.0000
## bwt_sd                  729.2143
## bwt_IQR                1073.0000
## bwt_mad                 834.7038
## bwt_tirst              2523.0000
## bwt_last               2495.0000
## bwt_4th                2594.0000

# frequency

bwt_tcf %>% count(bwt) %>% arrange(desc(n)) %>% head()

bwt	n
3062	5
2495	4
2920	4
2977	4
3651	4
1928	3

# use summary function

summary(bwt_tcf$bwt)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##     709    2414    2977    2945    3487    4990

# use describe function

describe(bwt_tcf$bwt) %>% t()

##                    X1
## vars        1.0000000
## n         189.0000000
## mean     2944.5873016
## sd        729.2142952
## median   2977.0000000
## trimmed  2961.7581699
## mad       834.7038000
## min       709.0000000
## max      4990.0000000
## range    4281.0000000
## skew       -0.2053370
## kurtosis   -0.1436834
## se         53.0425350

fivenum(bwt_tcf$bwt) #minimum, lower-hinge, median, upper-hinge, maximum

## [1]  709 2414 2977 3487 4990

boxplot.stats(bwt_tcf$bwt)

## $stats
## [1] 1021 2414 2977 3487 4990
## 
## $n
## [1] 189
## 
## $conf
## [1] 2853.682 3100.318
## 
## $out
## [1] 709

Whole data.frame

summary(bwt_tcf)

##       low         lowf     lowfl          age             lwt       
##  Min.   :0.0000   0:130   High:130   Min.   :14.00   Min.   : 80.0  
##  1st Qu.:0.0000   1: 59   Low : 59   1st Qu.:19.00   1st Qu.:110.0  
##  Median :0.0000                      Median :23.00   Median :121.0  
##  Mean   :0.3122                      Mean   :23.24   Mean   :129.8  
##  3rd Qu.:1.0000                      3rd Qu.:26.00   3rd Qu.:140.0  
##  Max.   :1.0000                      Max.   :45.00   Max.   :250.0  
##       race       racef  racegf  racegfl       smoke        smokef 
##  Min.   :1.000   1:96   0:96   white:96   Min.   :0.0000   0:115  
##  1st Qu.:1.000   2:26   1:26   black:26   1st Qu.:0.0000   1: 74  
##  Median :1.000   3:67   2:67   other:67   Median :0.0000          
##  Mean   :1.847                            Mean   :0.3915          
##  3rd Qu.:3.000                            3rd Qu.:1.0000          
##  Max.   :3.000                            Max.   :1.0000          
##        smokefl         ptl         ptlf    ptlgf  
##  Non_smoker:115   Min.   :0.0000   0:159   0:159  
##  Smoker    : 74   1st Qu.:0.0000   1: 24   1: 24  
##                   Median :0.0000   2:  5   2:  6  
##                   Mean   :0.1958   3:  1          
##                   3rd Qu.:0.0000                  
##                   Max.   :3.0000                  
##                     ptlgfl          ht          htf    
##  No_preterm_labor      :159   Min.   :0.00000   0:177  
##  One_preterm_labor     : 24   1st Qu.:0.00000   1: 12  
##  Two_plus_preterm_labor:  6   Median :0.00000          
##                               Mean   :0.06349          
##                               3rd Qu.:0.00000          
##                               Max.   :1.00000          
##               htfl           ui         uif    
##  No_hypertension:177   Min.   :0.0000   0:161  
##  Hypertension   : 12   1st Qu.:0.0000   1: 28  
##                        Median :0.0000          
##                        Mean   :0.1481          
##                        3rd Qu.:0.0000          
##                        Max.   :1.0000          
##                       uifl          ftv         ftvf    ftvgf  
##  No_uterine_irritability:161   Min.   :0.0000   0:100   0:100  
##  Uterine_irritability   : 28   1st Qu.:0.0000   1: 47   1: 47  
##                                Median :0.0000   2: 30   2: 42  
##                                Mean   :0.7937   3:  7          
##                                3rd Qu.:1.0000   4:  4          
##                                Max.   :6.0000   6:  1          
##                       ftvgfl         bwt       bwt_category
##  No_physician_visit      :100   Min.   : 709   low : 59    
##  One_physician_visit     : 47   1st Qu.:2414   high:130    
##  Two_plus_physician_visit: 42   Median :2977               
##                                 Mean   :2945               
##                                 3rd Qu.:3487               
##                                 Max.   :4990

lapply: Apply a Function over a List or Vector

#outputs a list

lapply(list(bwt_tcf$bwt, bwt_tcf$age, bwt_tcf$lwt), mean) 

## [[1]]
## [1] 2944.587
## 
## [[2]]
## [1] 23.2381
## 
## [[3]]
## [1] 129.8148

sapply: Apply a Function over a List or Vector

#outputs a vector

sapply(list(bwt_tcf$bwt, bwt_tcf$age, bwt_tcf$lwt), mean) 

## [1] 2944.5873   23.2381  129.8148

Grouping and summarizing

-for counting

-summarizing

Using tapply function

-Apply a Function Over a Ragged Array

# 1st dim(rows)=race, 2nd dim(cols)=smoke

tapply(bwt_tcf$bwt, list(bwt_tcf$race, bwt_tcf$smoke), mean)

##          0        1
## 1 3428.750 2826.846
## 2 2854.500 2504.000
## 3 2815.782 2757.167

#1st dim(rows)=race, 2nd dim(cols)=smoke, 3rd dim=ptl

tapply(bwt_tcf$bwt, list(bwt_tcf$race, bwt_tcf$smoke, bwt_tcf$ptl), mean) 

## , , 0
## 
##          0        1
## 1 3413.667 2902.225
## 2 2903.857 2474.500
## 3 2866.085 3065.750
## 
## , , 1
## 
##          0        1
## 1 4174.000 2452.000
## 2 2509.000 2622.000
## 3 2586.714 1766.667
## 
## , , 2
## 
##      0    1
## 1 3317 2601
## 2   NA   NA
## 3 2055 3260
## 
## , , 3
## 
##    0    1
## 1 NA 3637
## 2 NA   NA
## 3 NA   NA

Using aggregate function

# aggregate one or more continous variables by one factor


aggregate(bwt ~ race, data=bwt_tcf, mean)

race	bwt
1	3102.719
2	2719.692
3	2805.284

aggregate(cbind(bwt, lwt) ~ race, data=bwt_tcf, mean)

race	bwt	lwt
1	3102.719	132.0521
2	2719.692	146.8077
3	2805.284	120.0149

# aggregate one or more continous variables two factors


aggregate(bwt ~ race+smoke, data=bwt_tcf, mean)

race	smoke	bwt
1	0	3428.750
2	0	2854.500
3	0	2815.782
1	1	2826.846
2	1	2504.000
3	1	2757.167

aggregate(cbind(bwt, lwt) ~ race+smoke, data=bwt_tcf, mean)

race	smoke	bwt	lwt
1	0	3428.750	138.8409
2	0	2854.500	149.4375
3	0	2815.782	119.1455
1	1	2826.846	126.3077
2	1	2504.000	142.6000
3	1	2757.167	124.0000

# aggregate a number of continous variables by one or more factors

aggregate(. ~ race, data=bwt_tcf[c("bwt", "lwt", "age", "race")], mean) 

race	bwt	lwt	age
1	3102.719	132.0521	24.29167
2	2719.692	146.8077	21.53846
3	2805.284	120.0149	22.38806

aggregate(. ~ race+smoke, data=bwt_tcf[c("bwt", "lwt", "age", "race", "smoke")], mean)  

race	smoke	bwt	lwt	age
1	0	3428.750	138.8409	26.02273
2	0	2854.500	149.4375	19.93750
3	0	2815.782	119.1455	22.36364
1	1	2826.846	126.3077	22.82692
2	1	2504.000	142.6000	24.10000
3	1	2757.167	124.0000	22.50000

# aggregate a a continous variable by many grouping variable

aggregate(bwt ~ ., data = bwt_tcf[c("race","smoke", "ht", "bwt")], mean) 

race	smoke	ht	bwt
1	0	0	3436.395
2	0	0	2813.357
3	0	0	2874.824
1	1	0	2819.292
2	1	0	2656.111
3	1	0	2757.167
1	0	1	3100.000
2	0	1	3142.500
3	0	1	2063.000
1	1	1	2917.500
2	1	1	1135.000

# aggregate two continous variable by many grouping variable

aggregate(cbind(bwt, lwt) ~ ., data = bwt_tcf[c("race","smoke", "ht", "bwt", "lwt")], mean) 

race	smoke	ht	bwt	lwt
1	0	0	3436.395	139.2791
2	0	0	2813.357	143.4286
3	0	0	2874.824	118.9608
1	1	0	2819.292	121.9583
2	1	0	2656.111	137.6667
3	1	0	2757.167	124.0000
1	0	1	3100.000	120.0000
2	0	1	3142.500	191.5000
3	0	1	2063.000	121.5000
1	1	1	2917.500	178.5000
2	1	1	1135.000	187.0000

Using summaryBy function

-Opportunity to calculate multiple summary statistics for multiple variables

-Need to convert the tibble to data.frame for summaryBy analysis

library(doBy)

summary_fun <- function(x, ...){
  c(
  nobs=length(x),
  minm=min(x, na.rm=TRUE, ...), 
  maxm=max(x, na.rm=TRUE, ...), 
  qntl=quantile(x, na.rm=TRUE, ...), 
  iqr=IQR(x, na.rm=TRUE, ...), 
  MAD=mad(x, na.rm=TRUE, ...),
  avg=mean(x, na.rm=TRUE, ...),
  med=median(x, na.rm=TRUE, ...), 
  stdev=sd(x, na.rm=TRUE, ...),
  variance=var(x, na.rm=TRUE, ...)
  )
}

bwt_tcf_df <- as.data.frame(bwt_tcf) # tibble needs to be converted to a data.frame to work with the "doBy" package

# default function is 'mean'

summaryBy(bwt ~ race, data=bwt_tcf_df) 

race	bwt.mean
1	3102.719
2	2719.692
3	2805.284

summaryBy(bwt ~ race+smoke, data=bwt_tcf_df)

race	smoke	bwt.mean
1	0	3428.750
1	1	2826.846
2	0	2854.500
2	1	2504.000
3	0	2815.782
3	1	2757.167

summaryBy(cbind(bwt,age) ~ race+smoke, data=bwt_tcf_df) 

race	smoke	bwt.mean	age.mean
1	0	3428.750	26.02273
1	1	2826.846	22.82692
2	0	2854.500	19.93750
2	1	2504.000	24.10000
3	0	2815.782	22.36364
3	1	2757.167	22.50000

# default function is 'mean'

summaryBy(list(c("bwt", "age"), c("race", "smoke")), data=bwt_tcf_df )

race	smoke	bwt.mean	age.mean
1	0	3428.750	26.02273
1	1	2826.846	22.82692
2	0	2854.500	19.93750
2	1	2504.000	24.10000
3	0	2815.782	22.36364
3	1	2757.167	22.50000

# use "FUN" to define the functions 
# output according to the order of the functons listed

summaryBy(bwt ~ race + smoke, data=bwt_tcf_df, FUN=c(min, max, quantile, IQR, mean, median, sd, var))

race	smoke	bwt.FUN1	bwt.FUN2	bwt.FUN3	bwt.FUN4	bwt.FUN5	bwt.FUN6	bwt.FUN7	bwt.FUN8	bwt.FUN9	bwt.FUN10	bwt.FUN11	bwt.FUN12
1	0	1021	4990	1021	3062.00	3593.0	3873.00	4990	811.00	3428.750	3593.0	710.0989	504240.5
1	1	1790	4238	1790	2410.00	2775.5	3189.50	4238	779.50	2826.846	2775.5	626.4725	392467.8
2	0	1701	3860	1701	2480.75	2920.0	3331.25	3860	850.50	2854.500	2920.0	621.2543	385956.9
2	1	1135	3444	1135	2313.75	2381.0	2941.00	3444	627.25	2504.000	2381.0	637.0568	405841.3
3	0	1330	4054	1330	2313.00	2807.0	3253.00	4054	940.00	2815.782	2807.0	709.3493	503176.5
3	1	709	3572	709	2402.25	3146.5	3307.50	3572	905.25	2757.167	3146.5	810.0446	656172.3

summaryBy(cbind(bwt, age) ~ race + smoke, data=bwt_tcf_df, FUN=c(min, max, quantile, IQR, mean, median, sd, var))

race	smoke	bwt.FUN1	bwt.FUN2	bwt.FUN3	bwt.FUN4	bwt.FUN5	bwt.FUN6	bwt.FUN7	bwt.FUN8	bwt.FUN9	bwt.FUN10	bwt.FUN11	bwt.FUN12	age.FUN1	age.FUN2	age.FUN3	age.FUN4	age.FUN5	age.FUN6	age.FUN7	age.FUN8	age.FUN9	age.FUN10	age.FUN11	age.FUN12
1	0	1021	14	4990	45	1021	3062.00	3593.0	3873.00	4990	14	22.00	24.5	30.00	45	811.00	8.00	3428.750	26.02273	3593.0	24.5	710.0989	6.017373	504240.5	36.20877
1	1	1790	16	4238	33	1790	2410.00	2775.5	3189.50	4238	16	19.00	21.5	26.25	33	779.50	7.25	2826.846	22.82692	2775.5	21.5	626.4725	4.925807	392467.8	24.26357
2	0	1701	15	3860	27	1701	2480.75	2920.0	3331.25	3860	15	17.00	18.5	23.25	27	850.50	6.25	2854.500	19.93750	2920.0	18.5	621.2543	3.889623	385956.9	15.12917
2	1	1135	18	3444	35	1135	2313.75	2381.0	2941.00	3444	18	20.00	22.0	25.50	35	627.25	5.50	2504.000	24.10000	2381.0	22.0	637.0568	5.952591	405841.3	35.43333
3	0	1330	14	4054	33	1330	2313.00	2807.0	3253.00	4054	14	19.00	22.0	25.00	33	940.00	6.00	2815.782	22.36364	2807.0	22.0	709.3493	4.452896	503176.5	19.82828
3	1	709	14	3572	31	709	2402.25	3146.5	3307.50	3572	14	19.25	22.5	26.50	31	905.25	7.25	2757.167	22.50000	3146.5	22.5	810.0446	5.107926	656172.3	26.09091

# using custom function

summaryBy(bwt ~ race, data=bwt_tcf_df, FUN=summary_fun) 

race	bwt.nobs	bwt.minm	bwt.maxm	bwt.qntl.0%	bwt.qntl.25%	bwt.qntl.50%	bwt.qntl.75%	bwt.qntl.100%	bwt.iqr	bwt.MAD	bwt.avg	bwt.med	bwt.stdev	bwt.variance
1	96	1021	4990	1021	2584.75	3062	3651	4990	1066.25	867.3210	3102.719	3062	727.8861	529818.2
2	26	1135	3860	1135	2370.50	2849	3057	3860	686.50	693.8568	2719.692	2849	638.6839	407917.1
3	67	709	4054	709	2313.00	2835	3274	4054	961.00	693.8568	2805.284	2835	722.1944	521564.7

summaryBy(bwt ~ race + smoke, data=bwt_tcf_df, FUN=summary_fun)

race	smoke	bwt.nobs	bwt.minm	bwt.maxm	bwt.qntl.0%	bwt.qntl.25%	bwt.qntl.50%	bwt.qntl.75%	bwt.qntl.100%	bwt.iqr	bwt.MAD	bwt.avg	bwt.med	bwt.stdev	bwt.variance
1	0	44	1021	4990	1021	3062.00	3593.0	3873.00	4990	811.00	641.2245	3428.750	3593.0	710.0989	504240.5
1	1	52	1790	4238	1790	2410.00	2775.5	3189.50	4238	779.50	546.3381	2826.846	2775.5	626.4725	392467.8
2	0	16	1701	3860	1701	2480.75	2920.0	3331.25	3860	850.50	651.6027	2854.500	2920.0	621.2543	385956.9
2	1	10	1135	3444	1135	2313.75	2381.0	2941.00	3444	627.25	588.5922	2504.000	2381.0	637.0568	405841.3
3	0	55	1330	4054	1330	2313.00	2807.0	3253.00	4054	940.00	692.3742	2815.782	2807.0	709.3493	503176.5
3	1	12	709	3572	709	2402.25	3146.5	3307.50	3572	905.25	452.1930	2757.167	3146.5	810.0446	656172.3

# output to a tabular format

summaryBy(bwt ~ race + smoke, data=bwt_tcf_df, FUN=summary_fun) %>%
  t() %>%
  print(quote = FALSE, digits = 5)

##                       1         2         3         4         5         6
## race               1.00      1.00      2.00      2.00      3.00      3.00
## smoke              0.00      1.00      0.00      1.00      0.00      1.00
## bwt.nobs          44.00     52.00     16.00     10.00     55.00     12.00
## bwt.minm        1021.00   1790.00   1701.00   1135.00   1330.00    709.00
## bwt.maxm        4990.00   4238.00   3860.00   3444.00   4054.00   3572.00
## bwt.qntl.0%     1021.00   1790.00   1701.00   1135.00   1330.00    709.00
## bwt.qntl.25%    3062.00   2410.00   2480.75   2313.75   2313.00   2402.25
## bwt.qntl.50%    3593.00   2775.50   2920.00   2381.00   2807.00   3146.50
## bwt.qntl.75%    3873.00   3189.50   3331.25   2941.00   3253.00   3307.50
## bwt.qntl.100%   4990.00   4238.00   3860.00   3444.00   4054.00   3572.00
## bwt.iqr          811.00    779.50    850.50    627.25    940.00    905.25
## bwt.MAD          641.22    546.34    651.60    588.59    692.37    452.19
## bwt.avg         3428.75   2826.85   2854.50   2504.00   2815.78   2757.17
## bwt.med         3593.00   2775.50   2920.00   2381.00   2807.00   3146.50
## bwt.stdev        710.10    626.47    621.25    637.06    709.35    810.04
## bwt.variance  504240.47 392467.78 385956.93 405841.33 503176.47 656172.33

Using group_by function

# number of observations by group charcteristics

bwt_tcf %>% group_by(low, race) %>% count()

low	race	n
0	1	73
0	2	15
0	3	42
1	1	23
1	2	11
1	3	25

# number of observations by group charcteristics

bwt_tcf %>% group_by(low) %>% count(race)

low	race	n
0	1	73
0	2	15
0	3	42
1	1	23
1	2	11
1	3	25

# one grouping variable

bwt_tcf %>% 
  group_by(race) %>%
  summarise(no_of_obs=n(), 
            no_of_distinct_obs=n_distinct(bwt),
            
            bwt_avg=mean(bwt, na.rm = TRUE), 
            bwt_median=median(bwt, na.rm = TRUE),
            
            bwt_sd=sd(bwt, na.rm = TRUE),
            bwt_IQR=IQR(bwt, na.rm = TRUE),
            bwt_mad=mad(bwt, na.rm = TRUE),
            
            bwt_min=min(bwt, na.rm = TRUE),
            bwt_max=max(bwt, na.rm = TRUE),
            bwt_50th_quantile=quantile(bwt, 0.5, na.rm = TRUE),

            bwt_tirst=first(bwt),
            bwt_last=last(bwt),
            bwt_4th=nth(bwt, 4))

race	no_of_obs	no_of_distinct_obs	bwt_avg	bwt_median	bwt_sd	bwt_IQR	bwt_mad	bwt_min	bwt_max	bwt_50th_quantile	bwt_tirst	bwt_last	bwt_4th
1	96	80	3102.719	3062	727.8861	1066.25	867.3210	1021	4990	3062	2557	2495	2637
2	26	23	2719.692	2849	638.6839	686.50	693.8568	1135	3860	2849	2523	2495	2920
3	67	53	2805.284	2835	722.1944	961.00	693.8568	709	4054	2835	2551	2495	2722

# two grouping variables

bwt_tcf %>% 
  group_by(race, smoke) %>%
  summarise(no_of_obs=n(), 
            no_of_distinct_obs=n_distinct(bwt),
            
            bwt_avg=mean(bwt, na.rm = TRUE), 
            bwt_median=median(bwt, na.rm = TRUE),
            
            bwt_sd=sd(bwt, na.rm = TRUE),
            bwt_IQR=IQR(bwt, na.rm = TRUE),
            bwt_mad=mad(bwt, na.rm = TRUE),
            
            bwt_min=min(bwt, na.rm = TRUE),
            bwt_max=max(bwt, na.rm = TRUE),
            bwt_50th_quantile=quantile(bwt, 0.5, na.rm = TRUE),

            bwt_tirst=first(bwt),
            bwt_last=last(bwt),
            bwt_4th=nth(bwt, 4))

race	smoke	no_of_obs	no_of_distinct_obs	bwt_avg	bwt_median	bwt_sd	bwt_IQR	bwt_mad	bwt_min	bwt_max	bwt_50th_quantile	bwt_tirst	bwt_last	bwt_4th
1	0	44	41	3428.750	3593.0	710.0989	811.00	641.2245	1021	4990	3593.0	2637	2353	2877
1	1	52	44	2826.846	2775.5	626.4725	779.50	546.3381	1790	4238	2775.5	2557	2495	2663
2	0	16	15	2854.500	2920.0	621.2543	850.50	651.6027	1701	3860	2920.0	2523	2495	2920
2	1	10	9	2504.000	2381.0	637.0568	627.25	588.5922	1135	3444	2381.0	2920	2381	3444
3	0	55	45	2815.782	2807.0	709.3493	940.00	692.3742	1330	4054	2807.0	2551	2495	2722
3	1	12	12	2757.167	3146.5	810.0446	905.25	452.1930	709	3572	3146.5	3090	2495	3303

Basic graphics and additional descriptive statistics

bwt.df <- bwt.df %>% dplyr::select(bwt, age, lwt, everything())

head(bwt.df, n=3)

bwt	age	lwt	low	lowf	lowfl	race	racegr	racegf	racegfl	black	other	smoke	smokef	smokefl	ptl	ptlgr	ptlgf	ptlgfl	one_preterm_labor	twoOrMore_preterm_labor	ht	htf	htfl	ui	uif	uifl	ftv	ftvgr	ftvgf	ftvgfl	one_firstTimester_drVisit	twoOrMore_firstTimester_drVisit	bwt_category	bwt_cut	bwt_quantiles	bwt_tertile	bwt_tertile_order
2523	19	182	0	0	High	2	1	1	Black	1	0	0	0	Non-smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	1	1	Uterine_irritability	0	0	0	No_physician_visit	0	0	wt2	wt2	wt2	3	high
2551	33	155	0	0	High	3	2	2	Other	0	1	0	0	Non-smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	0	0	No_uterine_irritability	3	2	2	twoOrMore_firstTimester_drVisit	0	1	wt2	wt2	wt2	3	high
2557	20	105	0	0	High	1	0	0	White	0	0	1	1	Smoker	0	0	0	No_preterm_labor	0	0	0	0	No_hypertension	0	0	No_uterine_irritability	1	1	1	One_firstTimester_drVisit	1	0	wt2	wt2	wt2	3	high

Histograms, normal density curves and kernel density curves

par(mfrow=c(1,3))

for(i in 1:3) {
  hist(bwt.df[,i], prob=TRUE, main=names(bwt.df)[i])
  
  curve(dnorm(x, mean = mean(bwt.df[, i]), sd = sd(bwt.df[, i])), 
      col = "darkred", 
      lwd = 2,
      add = TRUE)
  
  lines(density(bwt.df[, i]), col = "blue", lwd = 2)
  
  rug(bwt.df[, i], col = "red", lwd = 2)
  
  }

Boxplots

par(mfrow=c(1,3))

for(i in 1:3) {
  boxplot(bwt.df[, i], main=names(bwt.df)[i])
}