Lecture 6 - Data Manipulation

Question 1

What is data wrangling?

Answer 1

Cleaning, transforming, and organizing data for analysis

Question 2

Name four data manipulation purposes.

Answer 2

1. Filtering ( Reorder the observations )
2. Variable creation
3. Calculating a set of summary statistics
4. Imputation

Question 3

How do you create new variables?

Answer 3

Apply functions to existing variables (e.g., BMI)

Question 4

What is imputation?

Answer 4

Adding or updating data to fill missing values

Question 5

List out 2 method for wrangling

Answer 5

1. dplyr
2. base R

Question 7

What scenario is recommended to use base R?

Answer 7

Minimal dependency or restricted environments

Question 8

What scenario is recommended to use dplyr?

Answer 8

1. Need clear, readable code, and optimisation

Question 9

How to apply a lookup table?

Answer 9

Use lookup_tbl[target_vector] to relabel values

Question 10

What is a tibble?

Answer 10

A modern data frame with enhanced printing and behavior

Question 11

How do tibbles print differently?

Answer 11

Show first 10 rows; neatly truncate columns

Question 12

Name a tibble creation function.

Answer 12

1. tibble()
2. as_tibble()

Question 13

How check if object is tibble? ( 2 )

Answer 13

1. is.tibble()
2. is.tbl()

Question 14

What verb selects columns?

Answer 14

1. select()

Question 15

What verb filters rows?

Answer 15

1. filter()

Subset Rows distinct(), slice()

Question 16

How to sort rows?

Answer 16

arrange() (use desc() for descending)

Question 17

How to add new columns?

Answer 17

mutate() (transmute() to keep only new ones)

Question 18

How to summarise data?

Answer 18

summarise() with summary functions

Question 19

What is the pipe operator?

Answer 19

1. %>% (magrittr)
2. |> (native R)

c(15, 20, 35) %>% mean()

%>% means then, which means we instantiate the vector then calculate its mean

Question 20

What does the pipe mean?

Answer 20

Pass output of one command to the next

Question 21

How to filter rows with base R and dplr?

Answer 21

1. base R -> subset()

data(flights, package="nycflights")
subset(flights, month==2 & day==1)
OR
flights |> subset(month==2 & day==1)

2. dply -> filter()

data(flights, package="nycflights")
filter(flights, month==2 & day==1)
OR
flights |> subset(month==2 & day==1)

Question 22

How to select base R and dplr?

Answer 22

1. []

data(flights, package="nycflights")
flights[, c("year","month","day")]

2. select()

data(flights, package="nycflights")
select(flights, year, month,day)

Question 23

How to group and summarise data with dplr?

Answer 23

1. group_by()

data(flights, package="nycflights")
flights |> group_by(month) -> flights_grouped_by_mth

2. summarise()

data(flights, package="nycflights")
flights |> group_by(month) |> summarise(mean(dep_delay, na.rm=T))

na.rm=T ignores the missing values

Question 24

How to create new columns from existing ones with dplyr?

Answer 24

1. mutate()

flights |>
mutate(
gain = arr_delay - delay,
air_time_hrs = air_time / 60,
speed = distance / air_time_hrs)

2. group() + mutate()

flights |>
group_by(carrier) %>%
mutate(
gain = arr_delay - delay,
air_time_hrs = air_time / 60,
speed = distance / air_time_hrs)

Question 25

Name five dplyr join types.

Answer 25

Inner Left Right Full Semi Anti