week14.rmd

---
title: "week14"
output: html_notebook
---

### 4.6.4

```{r}
hobby = c(
  'sport, reading, movie',
  'sport',
  'movie, sport, reading',
  'movie, Reading',
  'sport')
```

Who likes to read?

```{r}
hobby == "reading"
'sport, reading, movie' != "reading"
'sport, reading, movie' == 'sport, reading, movie'

hobby %in% "reading"
'sport, reading, movie' %in% 'sport, reading, movie'
```

```{r}
stringr::str_detect(
  string, # 'sport, reading, movie
  pattern # 'reading'
  )
```

```{r}
# any one likes to read
stringr::str_detect(
  hobby, # string
  "reading" # pattern
  ) # 4th is FALSE
stringr::str_detect(
  hobby, # string
  stringr::coll("reading", ignore_case = T) # pattern
  ) # 4th is TRUE
```

```{r}
table(hobby)
```

```{r}
stringr::str_split(
  string, 
  pattern
)
```

```{r}
stringr::str_split(
  hobby[[1]], 
  pattern = ", " # ", " != ","
) -> hobby1

stringr::str_split(
  hobby[[2]], 
  pattern = ", "
)

stringr::str_split(
  hobby[1:2], 
  pattern = ", "
) -> hobby12
class(hobby12)
# table(an atomic vector)

## turn a list into an atomic vector

table(unlist(hobby12))
```

```{r}
# Count for each hobby
table(hobby)

unlisted_hobbies <- {
  hobby |> 
    stringr::str_split(pattern = ", ") -> list_hobbies
  
  unlist(list_hobbies)
}
table(unlisted_hobbies)
```

```{r}
unlisted_hobbies
toupper(unlisted_hobbies) |> table()
tolower(unlisted_hobbies) |> table()
```

```{r}
# any one likes to read
pick_reading1 <- 
  stringr::str_detect(hobby, "reading") # 4th is FALSE
pick_reading2 <- 
  stringr::str_detect(hobby, stringr::coll("reading", ignore_case = T)) # 4th is TRUE
```

```{r}
pick_reading2
which(pick_reading2)
```

***

```{r}
df_dates =
  data.frame(
    year = c('2001','2001','2002','2001','2001'),
    month = c('4','10','1','1','4'),
    day = c('3','3','22','18','3')
  )
```

```{r}
df_dates$year
df_dates$month
df_dates$day

df_dates$data <- lubridate::ymd(c(
  "2001 4 3", "2001 10 3", "2002 1 22", "2001 1 18", "2001 4 3"
))

df_dates$date |> class()
```

```{r}
chr_dates <- paste(
  df_dates$year, 
  df_dates$month, 
  df_dates$day, 
  sep="-"
)
chr_dates
dates <- lubridate::ymd(chr_dates)
dates |> class()
```

#### factor vector

```{r}
students <-
  data.frame(
    major = c('economics','sociology','economics','sociology','sociology','finance','sociology','statistics','statistics','sociology'),
    year= c(4,1,4,3,1,4,4,2,1,3),
    credits= c(16,13,10,21,17,12,21,15,20,17)
  )
```

* school (學院)

```{r}
students$major |> unique()
students$major

school <- character(10)
school[c(1, 2, 3, 4, 5, 7, 10)] <- "social science"
school[c(6, 8, 9)] <- "business"
school
```

```{r}
school <- character(10)

{
  # For those whose major is economics or sociology, their school is social science. 
  pick_those = students$major %in% 
    c("economics", "sociology")
  school[pick_those] = "social science"
  
  # For those whose major is statistics or finance, their school is business
  pick_those = students$major %in% 
    c("statistics", "finance")
  school[pick_those] = "business"
}

school
```

```{r}
school = factor(students$major)

{
  levels(school) <- 
    c("social science","business","social science","business") # RHS and LHS must have the same length
  
  school
}

school
```

```{r}
students$year

students$year <- factor(students$year)
students$year

levels(students$year) <- c(
  "freshman", "sophemore", "junior", "senior"
)

students$year
```

numerical vecrtor into a categorical vector based on some interval definitions 

```{r}
students$credits
```

credits: <= 12 (light), 13-20 (normal), 20+ (heavy)

```{r}
workload <- character()
workload[c(3, 6)] <- "light"
workload[c(1, 2, 5, 8, 9, 10)] <- "normal"
workload[c(4, 7)] <- "heavy"
workload
```

```{r}
workload <- character(10)

pick_those = students$credits <= 12
workload[pick_those] = "light"

pick_those = (students$credits >= 13) | (students$credits <= 20)
workload[pick_those] = "normal"

pick_those = students$credits > 20
workload[pick_those] = "heavy"

workload
```

```{r}
workload <- factor(students$credits)

levels(workload) <- c(rep("light", 2),rep("normal", 5) , "heavy")

workload
```

* credits: <= 12 (light), 13-20 (normal), 20+ (heavy)

* light: 12, normal: 20, heavy: 30)

```{r}
maximalValues <- c(0, 12, 20, 30)
workload <- 
  cut(
    students$credits, # a numerical vector
    maximalValues # cut point
  )
workload # already a factor
students$credits

levels(workload) <- c("light", "normal", "heavy")

workload
students$credits
```

```{r}
credits2 <- c(0, students$credits)
credits2
cut(
  credits2, 
  maximalValues
)
```

```{r}
cut(
  credits2, 
  maximalValues, 
  include.lowest = T
) -> workload
levels(workload) <- c("light", "normal", "heavy")
workload
```


## 4.7 Summarise one vector

```{r}
dates <- c('2016-11-15','NA','NA','1997-05-07','1995-08-25','2002-09-20','NA','NA','NA','1995-07-16', '2011-06-22')
grades <- c(29,53,26,27,55,69,NA,NA,63,NA,56)
genders <- c('Male','Female','Male','Male','Female','Female',NA,'Male','Male','Female','Female')
majors <- c('economics','economics',NA,'economics','economics','economics','economics','statistics','law','economics','law')

dataSet <- 
  data.frame(
    date=dates,
    grade=grades, 
    gender=genders, 
    major=majors
  )

analysis <- list()
analysis$data <- dataSet
```

```{r}
View(dataSet)
```

```{r}
dataSet$date |> class()
dataSet$date |> lubridate::ymd() -> dataSet$date
dataSet$date |> class()

analysis$data <- dataSet
```

* analysis

```{r}
anyNA(dates)
dates |> is.na() |> sum() -> analysis$dates$na$sum
anyNA(grades)
grades |> is.na() |> sum() -> analysis$grades$na$sum

saveRDS(analysis, file = "analysis.Rds") # turn off computer
```

* only need to save ONE object

* a list gives me ability to use menu to manage my analysis process.

```{r}
analysis = readRDS("analysis.Rds")
analysis$grades$na$sum
analysis$data -> dataSet
```

```{r}
dates |> is.na() |> sum() -> a
grades |> is.na() |> sum() -> b
save(a, b, file="mywork.Rdata") # turn off computer
```

```{r}
load("mywork.Rdata")
```

***

Next day

```{r}
analysis = readRDS("analysis.Rds")
dates <- analysis$data$date
View(analysis$data)
```

```{r}
class(dates)
typeof(dates)

dates |> range() # minimal and maximal
dates |> range(na.rm=T) -> analysis$dates$range
analysis$dates$range
grades <- analysis$data$grade
grades |> range(na.rm=T) -> analysis$grades$range
analysis$grades$range

grades |> median(na.rm=T) -> analysis$grades$median
analysis$grades$median
grades |> mean(na.rm=T) -> analysis$grades$mean
analysis$grades$mean
```

```{r}
length(grades)
analysis$grades$na$sum
```

We have 11 students in the data set. For their grades, 3 are missing. The median grade is 54 (which means among 8 students with grades, 4 are lower than 54; 4 are higher than 54). The average grade is 47.25.

```{r}
genders <- analysis$data$gender
genders |> class()
genders |> factor() -> genders

genders |> levels() # only works for factor
```

```{r}
majors <- analysis$data$major
majors |> class()

majors |> unique() # returns a vector, data frame or array like x but with duplicate elements/rows removed.
```

```{r}
majors |> is.na() |> sum() -> analysis$majors$na$sum
majors |> factor() -> majors -> analysis$data$major

levels(majors)
```

* majors feature.

There are 11 students. One's major is missing. The remaining 10 students have majors from economic (7 students), law (2 students) and statistics(1 student).

```{r}
table(majors)

saveRDS(analysis, file="analysis.Rds")
```

```{r}
dataSet0 <- 
  data.frame(
    dates = c('2016-11-15','1997-05-07','NA','NA','1997-05-07','1995-08-25','2002-09-20','NA','NA','NA','1995-07-16', '2011-06-22', '2016-11-15'),
    grades = c(29,27, 53,26,27,55,69,NA,NA,63,NA,56, 29),
    genders = c('Male','Male', 'Female','Male','Male','Female','Female',NA,'Male','Male','Female','Female','Male'),
    majors = c('economics','economics', 'economics',NA,'economics','economics','economics','economics','statistics','law','economics','law','economics')
)
View(dataSet0)
```

```{r}
View(dataSet0)
duplicated(dataSet0)
whichIsDuplicated <- which(duplicated(dataSet0))
dataSet0[whichIsDuplicated, ]

dataSet0[whichIsDuplicated, ] -> dataSet0
View(dataSet0)
```

```{r}
dataSet0 <- unique(dataSet0)
View(dataSet0)
```

```{r}
# na is removed before table summarisation
genders |> table()
# preliminary summary should include NA summary
genders |> table(useNA = "always") -> analysis$genders$table
analysis$genders$table
```

```{r}
analysis = readRDS("analysis.Rds")
analysis$genders$table
```